Molecular and electronic structures by design: Tuning symmetrical and unsymmetrical linear trichromium chains

Article abstract of DOI:10.1021/ja049055h

The preparation, properties, and crystal structures of 12 trichromium extended metal atom chain (EMAC) compounds of the type Cr₃(L) ₄X₂ (L = equatorial ligands dipyridylamide (dpa) or di-4,4′-ethyl-2,2′-pyridylamide (depa), and X = axial ligands, e.g., halide or pseudohalide ions) with large variations in metal-metal distances are reported here. These complexes, which belong to a broad class of fundamentally interesting trinuclear molecules over which the electrons may or may not be delocalized, pose significant theoretical and experimental challenges which are dealt with in this report. Complexes with strongly donating axial or equatorial ligands tend to favor a symmetrical (D₄) molecular structure, while more weakly donating ligands give rise to unsymmetrical (C₄) structures; the physical properties of these two classes of compounds are discussed fully, and important comparisons with a reported DFT model of the electronic structures of the compounds are made.

Full text of DOI:10.1021/ja049055h

Published on Web 05/14/2004
Molecular and Electronic Structures by Design: Tuning
Symmetrical and Unsymmetrical Linear Trichromium Chains
John F. Berry, F. Albert Cotton,* Tongbu Lu,†,‡ Carlos A. Murillo,*
†
,†
,†
†
,§
†
Brian K. Roberts, and Xiaoping Wang
Contribution from the Department of Chemistry and Laboratory for Molecular Structure and
Bonding, P.O. Box 30012, Texas A&M UniVersity, College Station, Texas 77842-3012, School of
Chemistry & Chemical Engineering, Sun Yat-Sen UniVersity, Guangzhou 510275, P. R. China,
and Department of Chemistry, Florida State UniVersity, Tallahassee, Florida 32306
Received February 19, 2004; E-mail: cotton@tamu.edu; murillo@tamu.edu
Abstract: The preparation, properties, and crystal structures of 12 trichromium extended metal atom chain
3 4 2
(EMAC) compounds of the type Cr (L) X (L ) equatorial ligands dipyridylamide (dpa) or di-4,4′-ethyl-2,2′-
pyridylamide (depa), and X ) axial ligands, e.g., halide or pseudohalide ions) with large variations in metal-
metal distances are reported here. These complexes, which belong to a broad class of fundamentally
interesting trinuclear molecules over which the electrons may or may not be delocalized, pose significant
theoretical and experimental challenges which are dealt with in this report. Complexes with strongly donating
4
axial or equatorial ligands tend to favor a symmetrical (D ) molecular structure, while more weakly donating
ligands give rise to unsymmetrical (C ) structures; the physical properties of these two classes of compounds
4
are discussed fully, and important comparisons with a reported DFT model of the electronic structures of
the compounds are made.
1. Introduction
Scheme 1
Delocalized multicenter bonding in polyatomic systems poses
1
significant challenges to molecular quantum mechanics. Tri-
nuclear molecules or ions are the most fundamental examples
of a broad class of polyatomics which contain species such as
-
2
3
4
5
I3 , alkali or coinage metal trimers, carbon-rich chains such
6
7
8
+
as the allyl radical, O3, and the recently synthesized N5
cation, among others. Despite their simplicity, these entities
tend not to be structurally rigid. For example, systems with three
electrons (Scheme 1) can be either localized having a doublet
ground state with unequal (I) or equal (II) interatomic distances,
or delocalized (III) having equal interatomic distances. In the
latter, the three electrons are in orbitals spanning all three atoms,
and these can be either paired or unpaired.
9
10
†
Texas A&M University.
Sun Yat-Sen University.
Florida State University.
‡
§
(
1) (a) Munzarov a´ , M. L.; Hoffmann, R. J. Am. Chem. Soc. 2002, 124, 4787.
(5) (a) Brown, L. D.; Lipscomb, W. N. J. Am. Chem. Soc. 1977, 99, 3968. (b)
Siddarth, P.; Gopinathan, M. S. J. Am. Chem. Soc. 1988, 110, 96. (c) Ohno,
M.; Zakrzewski, V. G.; Ortiz, J. V.; von Niessen, W. J. Chem. Phys. 1997,
106, 3258. (d) Ortiz, J. V.; Zakzewski, V. G. J. Chem. Phys. 1994, 100,
6614. (e) Hiberty, P. C.; Ohanessian, G.; Shaik, S. S.; Flament, J. P. Pure
Appl. Chem. 1993, 65, 35.
(6) (a) Shaik, S. S.; Hiberty, P. C.; Ohanessian, G.; Lefour, J. M. J. Phys.
Chem. 1988, 92, 5086. (b) Shaik, S. S.; Hiberty, P. C.; Ohanessian, G.;
Lefour, J. M. NouV. J. Chim. 1985, 9, 385. (c) Shaik, S. S.; Hiberty, P. C.;
Lefour, J. M.; Ohanessian, G. J. Am. Chem. Soc. 1987, 109, 363. (d)
Berthier, G. Can. J. Chem. 1985, 63, 1681.
(
b) Macchi, P.; Sironi, A. Coord. Chem. ReV. 2003, 238, 383. (c) Harcourt,
R. D. J. Phys. Chem. A 1999, 103, 4293. (d) Harcourt, R. D. J. Phys. Chem.
A 2003, 107, 11260. (e) Ramsden, C. A. Chem. Soc. ReV. 1994, 111. (f)
Vela, A.; Merino, G. ReV. Mod. Quantum Chem. 2002, 2, 1140. (g) Rasul,
G.; Olah, G. A.; Prakash, G. K. S. Inorg. Chem. 2003, 42, 8059. (h) Kar,
T.; AÄ ngy a´ n, J. G.; Sannigrahi, A. B. J. Phys. Chem. A 2000, 104, 9953. (i)
Alkorta, I.; Abboud, J. L. M.; Quintanilla, E.; D a´ valos, J. Z. J. Phys. Org.
Chem. 2003, 16, 564. (j) Harcourt, R. D. J. Mol. Struct. (THEOCHEM)
2
002, 617, 177. (k) Sannigrahi, A. B.; Kar, T. J. Mol. Struct. (THEOCHEM)
000, 496, 1. (l) Shaik, S.; Shurki, A.; Danovich, D.; Hiberty, P. C. Chem.
2
ReV. 2001, 101, 1501. (m) Jug, K.; Hiberty, P. C.; Shaik, S. Chem. ReV.
(7) (a) Laing, M. Struct. Chem. 1995, 6, 397. (b) Laing, M. Educ. Chem. 1996,
46.
(8) Christe, K. O.; Wilson, W. W.; Sheehy, J. A.; Boatz, J. A. Angew. Chem.,
Int. Ed. 1999, 38, 2004.
2
001, 101, 1477.
2) (a) Svensson, P. H.; Kloo, L. J. Chem. Soc., Dalton Trans. 2000, 14, 2449.
b) Margulis, C. J.; Coker, D. F.; Lynden-Bell, R. M. Chem. Phys. Lett.
(
(
2
3
2
001, 341, 557. (c) Koslowski, T.; V o¨ hringer, P. Chem. Phys. Lett. 2001,
42, 141. (d) Kloo, L.; Rosdahl, J.; Svensson, P. H. Eur. J. Inorg. Chem.
002, 1203. (e) Lin, Z.; Hall, M. B. Polyhedron 1993, 12, 1499. (f) Vala,
(9) (a) Netzloff, H. M.; Gordon, M. S.; Chroste, K.; Wilson, W. W.; Vij, A.;
Vij, V.; Boatz, J. A. J. Phys. Chem. A 2003, 107, 6638. (b) Fau, S.; Wilson,
K. J.; Bartlett, R. J. J. Phys. Chem. A 2002, 106, 4639. (c) Kerkines, I. S.
K.; Papakondylis, A.; Mavridis, A. J. Phys. Chem. A 2002, 106, 4453. (d)
Wang, L. J.; Qian, S. L.; Warburton, P.; Mezey, P. G. J. Phys. Chem. A
2002, 106, 1872. (e) Ponec, R.; Roithov a´ , J.; Giron e´ s, X.; Jug, K. J. Mol.
Struct. (THEOCHEM) 2001, 545, 255. (f) Harcourt, R. D.; Klap o¨ tke, T.
M. Z. Naturforsch. 2002, 57, 983. (g) Harcourt, R. D.; Klap o¨ tke, T. M. Z.
Naturforsch. 2003, 58, 121.
J.; Kosloff, R.; Harvey, J. N. J. Chem. Phys. 2001, 114, 7413. (g) Margulis,
C. J.; Coker, D. F.; Lynden-Bell, R. M. J. Chem. Phys. 2001, 114, 367. (h)
Sato, H.; Hirata, F.; Myers, A. B. J. Phys. Chem. A 1998, 102, 2065. (i)
Zhang, F. S.; Lyden-Bell, R. M. Phys. ReV. Lett. 2003, 90, 185505-1. (j)
Gorinchoi, N. N.; Cimpoesu, F.; Bersuker, I. B. J. Mol. Struct. (THEOCHEM)
2
000, 530, 281.
(
3) (a) Martins, J. L.; Car, R.; Buttet, J. J. Chem. Phys. 1983, 78, 5646. (b)
Malrieu, J. P. NouV. J. Chim. 1986, 10, 61.
(10) (a) Schuh, W.; Braunstein, P.; B e´ nard, M.; Rohmer, M.-M.; Welter, R.
Angew. Chem., Int. Ed. 2003, 42, 2161. (b) Shaik, S. S.; Bar, R. Inorg.
Chem. 1983, 22, 735. (c) Xu, L.; Ugrinov, A.; Sevov, S. C. J. Am. Chem.
Soc. 2001, 123, 4091.
(4) Berthier, G.; Barthelat, J. C.; Dangeard, I.; Tao, Y. J. Phys.-Chim. Biol.
1
987, 84, 677.
7082
9
J. AM. CHEM. SOC. 2004, 126, 7082-7096
10.1021/ja049055h CCC: $27.50 © 2004 American Chemical Society

Molecular and Electronic Structures by Design
A R T I C L E S
Polynuclear systems having metal-metal multiple bonds have
unsymmetrical-1) which are different only by the length
2
1
also been of considerable interest both to experimentalists and
of one or more bonds. The analogous tricobalt compound,
Co3(dpa)4Cl2, unequivocally exists in two different and well-
characterized forms, s-Co3(dpa)4Cl2 and u-Co3(dpa)4Cl2, which
11
theorists. The simpler dinuclear systems have been extensively
studied, and in many cases the electronic structure is well
1
1
22
understood, an exception being those with quadruply bonded
fit the description of bond stretch isomers. For chromium
Cr2 units. Many compounds of this type have been prepared,12
and they exhibit a wide variation of Cr-Cr bond distances
ranging from 1.83 to ∼2.6 Å. These are essentially diamag-
netic13 and are described as having a formal Cr-Cr quadruple
bond. The reason for this remarkable variability of Cr-Cr bond
lengths is that axial donors L destabilize the Cr-Cr σ bond
4+
complexes of the type Cr3(dpa)4XY, symmetrical chains were
reported when X ) Y, but when X * Y the Cr-Cr distances
6
+
become so different that the Cr3 chain could be described as
4+
consisting of a diamagnetic Cr2 quadruply bonded unit
tethered to a high-spin Cr(II) ion at a nonbonding distance (see
23
Scheme 3). Both symmetrical and unsymmetrical compounds
1
4
15
through either σ or π interactions. These chromium com-
pounds have presented major challenges from a theoretical
standpoint,16 and reliable calculations on Cr2 complexes have
Scheme 3
4+
1
7
been difficult to obtain due to electron correlation effects.
4+
Nevertheless, the description of Cr2 compounds as having a
2
4 2
σ π δ configuration has been helpful, if not a totally correct
approximation,1 and it has permitted an explanation of the
geometry and electronic structure and those of a recently
1
synthesized Cr25 cationic unit.
+
18
have S ) 2 ground states. This is in contrast to Co3(dpa)4Cl2
In 1997, this laboratory reported the first compounds with
metal-metal bonded units forming extended metal atom
chains (EMACs) of three Cr(II) atoms, the prototype being
Cr3(dpa)4Cl2 (1, dpa is the anion of di-2,2′-pyridylamine, see
Scheme 2).19
for which the symmetrical and unsymmetrical “isomers” have
22
qualitatively different magnetic properties inter alia.
To help explain the experimental findings, DFT calculations
have been carried out by B e´ nard, Rohmer, and co-workers on
24
the parent complex Cr3(dpa)4Cl2. According to these calcula-
tions, there is a symmetrical ground state with dCr-Cr ) 2.350
Å. This was rationalized in terms of a 3-center-3-electron bond
consisting of a filled 3-center σ bonding orbital and a half-
filled σ nonbonding orbital. The remaining 9 Cr-based electrons
are localized on the three Cr atoms in orbitals of π and δ
symmetry and couple antiferromagnetically and very strongly
to produce the observed S ) 2 ground state shown in Scheme
4. The relative energies of the σ and σnb orbitals are influenced
Scheme 2
In subsequent work with trichromium complexes of this
type, we found that in some cases the Cr-Cr distances were
equal, but in others a rather unsymmetrical Cr3 chain
2
0
Scheme 4
exists. This prompted us to consider 1 as existing as a pair of
bond stretch isomers: two compounds (symmetrical-1 and
(
11) Cotton, F. A.; Walton, R. A. Multiple Bonds Between Metal Atoms, 2nd
ed.; Oxford University Press: New York, 1993.
(
12) (a) Cotton, F. A.; Murillo, C. A.; Pascual, I. Inorg. Chem. 1999, 38, 2182.
(b) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Schooler, P. J. Chem.
Soc., Dalton Trans. 2000, 2007. (c) Hao, S.; Gambarotta, S.; Bensimon,
C.; Edema, J. J. H. Inorg. Chim. Acta 1993, 213, 65. (d) Cotton, F. A.;
Daniels, L. M.; Murillo, C. A.; Schooler, P. J. Chem. Soc., Dalton Trans.
2
000, 2001. (e) Carlton-Day, K. M.; Eglin, J. L.; Lin, C.; Smith, L. T.;
Staples, R. J.; Wipf, D. O. Polyhedron 1999, 18, 817. (f) Dionne, M.; Hao,
S.; Gambarotta, S. Can. J. Chem. 1995, 73, 1126. (g) Cotton, F. A.;
Timmons, D. J. Polyhedron 1998, 17, 179. (h) Mashima, K.; Tanaka, M.;
Tani, K.; Nakamura, A.; Tadeka, S.; Mori, W.; Yamaguchi, K. J. Am. Chem.
Soc. 1997, 119, 4307.
by destabilizing interactions with the axial ligands.
Although a very unsymmetrical structure for Cr3(dpa)4Cl2
4
+
(
13) In some cases, weak paramagnetism is observed for Cr
2
species due to
(
e.g., with ∆dCr-Cr ) 0.679 Å) was calculated to lie >4 kcal
the partial population of an excited triplet state, although the ground state
has S ) 0. See: Cotton, F. A.; Chen, H.; Daniels, L. M.; Feng, X. J. Am.
Chem. Soc. 1992, 114, 8980.
-
1
mol above the symmetrical ground state, it was postulated
that an unsymmetrical set of axial ligands could stabilize such
a state, which rationalizes the very unsymmetrical structures of
(
(
(
14) Cotton, F. A.; Hillard, E. A.; Murillo, C. A.; Zhou, H.-C. J. Am. Chem.
Soc. 2000, 122, 416.
15) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Pascual, I.; Zhou, H.-C. J.
Am. Chem. Soc. 1999, 121, 6856.
25
Cr3(dpa)4(BF4)Cl and Cr3(dpa)4(PF6)Cl. Intermediate geom-
-1
16) (a) Ponec, R.; Yuzhakov, G.; Carb o´ -Dorca, R. J. Comput. Chem. 2003,
etries with ∆dCr-Cr ≈ 0.1 Å were calculated to lie <1 kcal mol
2
4, 1829. (b) Lichtenberger, D. L.; Lynn, M. A.; Chisholm, M. H. J. Am.
24
above the ground state, and thus the energy required to move
the central Cr atom toward one of the two terminal Cr atoms is
quite small, and about the same order of magnitude as kT.
Chem. Soc. 1999, 121, 12167. (c) Nishimo, M.; Yamanaka, S.; Yoshioka,
Y.; Yamaguchi, K. J. Phys. Chem. A 1997, 101, 705. (d) Nishino, M.;
Yoshioka, Y.; Yamaguchi, K.; Mashima, K.; Tani, K.; Nakamura, A. Bull.
Chem. Soc. Jpn. 1998, 71, 99. (e) Andersson, K.; Bauschlicher, C. W., Jr.;
Persson, B. J.; Roos, B. O. Chem. Phys. Lett. 1996, 257, 238.
(
(
17) Hall, M. B. Polyhedron 1987, 6, 679.
(21) Parkin, G. Acc. Chem. Res. 1992, 25, 455.
18) (a) Cotton, F. A.; Daniels, L. M.; Huang, P.; Murillo, C. A. Inorg. Chem.
(22) Cl e´ rac, R.; Cotton, F. A.; Daniels, L. M.; Dunbar, K. R.; Kirschbaum, K.;
Murillo, C. A.; Pinkerton, A. A.; Schultz, A. J.; Wang, X. J. Am. Chem.
Soc. 2000, 122, 6226.
2
002, 41, 317. (b) Cotton, F. A.; Dalal, N. S.; Hillard, E. A.; Huang, P.;
Murillo, C. A.; Ramsey, C. M. Inorg. Chem. 2003, 42, 1388.
19) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Pascual, I. J. Am. Chem.
Soc. 1997, 119, 10223.
(
(
(23) Cl e´ rac, R.; Cotton, F. A.; Daniels, L. M.; Dunbar, K. R.; Murillo, C. A.;
Pascual, I. Inorg. Chem. 2000, 39, 748.
(24) Benbellat, N.; Rohmer, M.-M.; B e´ nard, M. Chem. Commun. 2001, 2368.
(25) Rohmer, M.-M.; B e´ nard, M. J. Cluster Sci. 2002, 13, 333.
20) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Pascual, I. Inorg. Chem.
Commun. 1998, 1, 1.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7083

A R T I C L E S
Berry et al.
Scheme 5
at 80 °C under vacuum prior to use. Ferrocenium triiodide was prepared
by mixing hexanes solutions of appropriate amounts of ferrocene and
29
2
I .
Physical Measurements. The IR spectra were taken on a Perkin-
Elmer 16PC FTIR spectrometer using KBr pellets. Cyclic voltammo-
grams were taken on a CH Instruments electrochemical analyzer using
4 6
dichloromethane solutions with 1 M NBu PF and 0.1 mM analyte.
The electrodes were as follows: Pt disk (working), Pt wire (auxiliary),
and Ag/AgCl (reference). Elemental analyses were carried out by
Canadian Microanalytical Services in British Columbia, Canada, except
for 2, which was done at the microanalytical laboratory at Sun
Yat-Sen University in Guangzhou, P. R. China. Samples were vacuum-
dried prior to elemental analyses to remove the interstitial solvent
molecules of the crystals. Note that this procedure resulted in partial
removal of the axial acetonitrile ligands from 4. Magnetic susceptibility
measurements were made on crushed crystalline samples from which
the interstitial solvents had been removed by vacuum. The samples
were carefully weighed and placed in a plastic bag, which was mounted
in a drinking straw and then placed inside a Quantum Design SQUID
magnetometer MPMS-XL. Variable temperature data were collected
from 2 to 300 K at a field of 1000 G, and the data were corrected
We began this work to provide some insight into the factors
which are ultimately responsible for the appearance of a
symmetrical versus an unsymmetrical Cr36+ chain in the solid
state and to see if the molecular geometry could be tuned by
changing the electronic effects of the axial and equatorial
ligands. We also wanted to see if it was possible to find more
conclusive evidence either for or against 1 forming bond stretch
isomers. During the course of this work, we found it necessary
to reexamine some of our previously reported results on
2
3
Cr3(dpa)4Cl2, as well as to synthesize new compounds with
various axial donor ligands and Cr36+ and Cr3 compounds
stabilized by the substituted depa (i.e., the anion of di-4,4′-ethyl-
7+
1
empirically for diamagnetism of the sample and the holder. H NMR
2
,2′-pyridylamine) equatorial ligand (see Scheme 5).
spectra were obtained on a VXR-300 NMR spectrometer. Mass
spectrometry data (electrospray ionization) were recorded at the
Laboratory for Biological Mass Spectrometry at Texas A&M Univer-
sity, using an MDS Series Qstar Pulsar with a spray voltage of 5 keV.
Visible spectra were obtained on either a Shimadzu UV-2501 PC UV-
vis spectrophotometer or a Cary 17D spectrophotometer. Electronic
reflectance spectra were obtained in the region from 400 to 1000 nm
on a HP 845x UV-visible system with a Labsphere RSA-HP-8453
reflectance accessory.
The compounds are as follows: Cr3(dpa)4Cl2, 1; Cr3(depa)4-
Cl2, 2; [Cr3(depa)4Cl2]I3, 3; [Cr3(dpa)4(NCCH3)2](PF6)2, 4; Cr3-
dpa)4I2, 5; Cr3(dpa)4Br2, 6; Cr3(dpa)4(NO3)2, 7; Cr3(dpa)4(BF4)2,
(
8
1
; Cr3(dpa)4(CN)2, 9; Cr3(dpa)4(NCS)2, 10; Cr3(dpa)4(NCO)2,
1; and [Cr3(dpa)4(NCO)F]PF6, 12. Further work on Cr3(dpa)4-
(CCPh)2, 13, is also reported.
The results of this work are relevant to the areas of metal-
metal bonding11 and molecular electronics, since changes in
molecular geometries that will either enhance or diminish
delocalized metal-metal bonding are of utmost importance to
the design of useful molecular switches. The possibility of
switching the molecular structure of a trichromium compound
such as 1 from symmetrical to unsymmetrical through an outside
source such as an applied potential comes to mind. Moreover,
compounds which can behave electronically as switches are
useless if they cannot be integrated into a circuit. So, we report
our exploration of the chemistry of Cr3(dpa)4X2 compounds and
methods for varying the X anions which should be applicable
to a variety of X ligands.
26
Syntheses. Cr
dissolved in 25 mL of THF. The solution was cooled in a dry ice/
acetone bath, and 1.3 mL of 1.6 M CH Li in ether was added slowly.
3 4 2
(depa) Cl (2): Hdepa (0.45 g, 2.0 mmol) was
3
The yellow solution was allowed to warm to room temperature, and
then was transferred via cannula to a reaction flask containing anhydrous
CrCl (0.19 g, 1.5 mmol). The red mixture was stirred at room
2
temperature for 1 h, and then refluxed for 3 h, giving a deep green
mixture. The solvent was removed under vacuum, the remaining green-
brown solid was extracted with toluene (2 × 8 mL), and the solution
was then layered with hexanes. After 1 week, deep green crystals of
Cr
with hexanes several times, and dried under vacuum. Yield: 0.272 g,
8%. Anal. Calcd for C59Cl Cr 12: C, 60.33; H, 6.04; N, 14.30.
Found: C, 60.15; H, 5.88; N, 13.99. Mass spectrum, ESI + (m/z): 1130
. IR (KBr, cm-1): 3751 w, 3448 br, w, 3056 w, 2964 m, 1611 vs,
3
(depa)
4
Cl
2
‚0.5C
6
H14 formed. The crystals were collected, washed
4
2
3 71
H N
2
. Experimental Section
M
+
1
9
532 m, 1473 s, 1424 vs, 1300 m, 1225 w, 1179 m, 1061 m, 1016 s,
General. All manipulations were carried out under dry nitrogen using
-1
2 2
28 m, 809 s, 547 w, 436 w. Vis (CH Cl solution: λ, nm (ꢀ, M
standard Schlenk techniques. THF, ether, dichloromethane, hexanes,
toluene, and benzene were purified using a Glass Contour system. All
other solvents were distilled in a nitrogen atmosphere over appropriate
drying agents prior to use, and reagents available from commercial
-1
cm )): 664 (900), 608 (1000), 511 (2000), 436 (4000).
[Cr
3
(depa)
4
Cl
2 3
]I (3): To a mixture of 80 mg (0.071 mmol) of
Cr
3
(depa)
4
Cl
2
and 40 mg (0.071 mmol) of ferrocenium triiodide was
added 10 mL of dichloromethane. The resulting dark green solution
was stirred for 2 h, and then 30 mL of hexanes was added, precipitating
a dark green solid and leaving a pale yellow solution which was
removed by filtration. The solid was washed with 10 mL of hexanes
and redissolved in 10 mL of dichloromethane. This was filtered into a
crystallization tube where it was layered with hexanes. Large dark green
sources were used as received unless otherwise specified. Cr
3
(dpa)
4
-
2
3
Cl
synthesized as published. Commercial grade anhydrous CrCl
refluxed in Me SiCl prior to use. Thallium hexafluorophosphate, KCN,
NaI, KBr, TlNO , NaNCO, and KSCN were dried overnight under
dynamic vacuum at 80 °C prior to use. Cr (dpa) (CCPh) , 13, was
prepared according to a published procedure. Thallium tetrafluorobo-
rate was prepared from TlCO and HBF in ether and dried overnight
2
, 1, was prepared as reported previously. The ligand Hdepa was
2
7
2
was
3
3
3
4
2
crystals of [Cr
collected, washed with hexanes, and dried under vacuum. Yield: 90
mg, 84%. Anal. Calcd for Cr : C, 44.46; H, 4.23; N,
1.11. Found: C, 43.97; H, 4.24; N, 10.73. Mass spectrum, ESI +
3 4 2 3
(depa) Cl ]I formed in about a week. These were
28
3
4
3
56 12 2 3
C N H64Cl I
1
(
(
(
26) For a recent review of molecular electronics, see: Carroll, R. L.; Gorman,
+
-
-1
(m/z): 1130 [Cr
3
(depa)
4
Cl
2
] , ESI - (m/z): 381 I . IR (KBr, cm ):
3
C. B. Angew. Chem., Int. Ed. 2002, 41, 4378.
27) Berry, J. F.; Cotton, F. A.; Lu, T.; Murillo, C. A.; Wang, X. Inorg. Chem.
3448 w, br, 2965 w, 2930 w, 1615 s, 1529 w, 1471 m, 1423 vs, 1346
2
003, 42, 3595.
28) Berry, J. F.; Cotton, F. A.; Murillo, C. A.; Roberts, B. K. Inorg. Chem.
004, 43, 2277.
2
(29) Neuse, E. W.; Loonat, M. S. J. Organomet. Chem. 1985, 286, 329.
7084 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
A R T I C L E S
m, 1226 w, 1181 w, 1058 w, 1016 m, 930 m, 859 w, 816 m, 739 w,
1264 w, 1156 m, 1122 m, 1110 m, 1085 m, 1019 m, 955 w, 928 m,
883 w, 860 w, 805 w, 766 m, 749 w, 649 w, 537 w, 517 w, 436 w,
421 w.
5
51 w, 438 w.
Cr (dpa) (NCMe)
mmol) of Cr (dpa) Cl
5 mL of acetonitrile. After the resulting green mixture was stirred for
h, a fine white precipitate was observed. The mixture was filtered
[
3
4
2
](PF
6
)
2
(4): To a flask containing 858 mg (0.946
was added
3
4
2
and 661 mg (1.89 mmol) of TlPF
6
3 4 2
Cr (dpa) (CN) (9): In a drybox, 0.098 g (0.081 mmol) of 4 was
added to a Schlenk flask. Dried and powdered KCN (0.011 g, 0.17
mmol) was weighed in air and added to a second flask, followed by
2
4
through Celite, giving a clear dark olive green solution, which was
layered with ether. After diffusion of the ether into the acetonitrile
evacuation and purging with N
added while stirring. The solution of KCN was transferred to the Cr
2
. To each flask, 10 mL of MeOH was
3
solution, large green block-shaped crystals of [Cr
PF ‚2CH CN formed. These were collected, washed with hexanes,
and dried under vacuum. Yield: 824 mg, 72%. Anal. Calcd for
12: C, 42.93; H, 2.94; N, 15.26. Found: C, 42.44;
H, 3.29; N, 15.54. IR (KBr, cm ): 3448 w, br, 2268 w (CtN), 1608
s, 1551 m, 1468 vs, 1429 vs, 1362 s, 1314 m, 1284 m, 1244 w, 1159
3
(dpa)
4
(NCCH
3
)
2
]-
flask via cannula. The mixture was stirred for 5 min, after which the
solvent was removed under vacuum while heating. The product was
extracted from the dried residue with 10 mL of CH Cl . The extract
2 2
was filtered over Celite into a Schlenk tube, and the solution was then
carefully layered with hexanes. Large dark green block-shaped crystals
were observed the following day. Yield: 37 mg, 51%. Anal. Calcd for
(
6
)
2
3
C
41
H
33.5
N
12.5Cr
3
P
2
F
-
1
s, 1114 w, 1058 w, 1020 m, 840 vs (PF
m, 517 w, 437 w. Vis (CH
CN solution: λ, nm (ꢀ, M- cm )): 679
2000), 561 (1000), 515 (sh, 2000), 435 (sh, 4000).
Cr (dpa) (5): In a drybox, 0.105 g (0.087 mmol) of 4 was added
to a Schlenk flask. To a second flask was added 0.028 g (0.19 mmol)
of NaI that had been powdered and weighed in air. The flask was
6
), 766 s, 741 m, 650 w, 557
Cr
3
C
43.5
H
35
N
14Cl
3
: C, 51.41; H, 3.47; N, 19.30. Found: C, 51.79; H,
1
-1
-1
3
3.60; N, 19.14. IR (KBr, cm ): 3448 w, br, 3028 w, 2965 w, 2370 w,
2344 w, 2096 w, CtN, 2063 w, 1606 s, 1595 s, 1546 w, 1467 vs,
1428 vs, 1364 s, 1314 m, 1282 m, 1264 m, 1153 m, 1109 m, 1015 m,
879 w, 859 w, 803 w, 765 m, 735 m, 645 w, 537 w, 517 w, 474 w,
(
3
4 2
I
+
435 w. Mass spectrum, ESI + (m/z): 888 (M - H) .
2
evacuated and purged with N , and 10 mL of freshly distilled MeOH
3 4 2
Cr (dpa) (NCS) (10): In the drybox, 0.076 g (0.063 mmol) of 4
was added to each flask, and then the NaI solution was transferred via
cannula to the flask containing 4. The color darkened, and a crystalline
precipitate formed. The mixture was stirred for 1 h. The green solid
was collected by filtration and washed with diethyl ether (2 × 20 mL).
was added to a Schlenk flask. To a second flask was added 0.013 g
(0.13 mmol) of dried and powdered KSCN. Next, 10 mL of MeOH
was added to each flask while stirring. The contents of the KSCN flask
were transferred to the other flask via cannula. After the mixture was
stirred for ∼15 min, a green precipitate was observed. This was
2 2
The solid was dissolved in 12 mL of CH Cl and filtered into a Schlenk
tube. The green solution was layered with hexanes, and block-shaped
crystals were observed the following day. These were collected, washed
with hexanes, and dried under vacuum. Yield: 70 mg, 74%. Anal. Calcd
collected by filtration, washed with Et
2
O (2 × 10 mL), and dissolved
in toluene (12 mL). The resulting green mixture was filtered over Celite
into a Schlenk tube, and the solution was layered with hexanes. Crystals
grew as long (10 mm +) but very thin needles within a week. Yield:
3 40 32 12 2
for Cr C H N I : C, 44.05; H, 2.96; N, 15.41. Found: C, 44.63; H,
-
1
2
2
1
8
.86; N, 15.92. IR (KBr, cm ): 3448 w, br, 3096 w, 3067 w, 3024 w,
3 42 36 14 2 2
53 mg, 88%. Anal. Calcd for Cr C H N O S : C, 51.01; H, 3.67;
-
1
370 w, 2344 w, 1606 s, 1595 s, 1547 w, 1466 vs, 1426 vs, 1362 s,
314 m, 1282 w, 1265 w, 1152 m, 1110 w, 1055 w, 1012 w, 916 w,
78 w, 860 w, 763 m, 734 m, 643 w, 537 w, 515 w, 434 w, 420 w.
N, 19.37. Found: C, 50.83; H, 3.56; N, 19.37. IR (KBr, cm ): 3447
w, br, 3071 w, 3028 w, 2963 w, 2368 w, 2342 w, 2028 vs, CdN,
1597 s, 1547 w, 1463 vs, 1425 vs, 1367 s, 1311 m, 1279 w, 1155 m,
Cr
3
(dpa)
4
Br
2
(6): This was synthesized similarly to 5. IR (KBr,
1108 w, 1055 w, 1015 m, 916 w, 879 w, 857 w, 802 w, 764 m, 738 w,
cm- ): 3027 w, 2963 w, 1607 s, 1595 s, 1547 m, 1468 vs, 1429 vs,
1
-1
2 2
645 w, 538 w, 518 w, 434 w. Vis (CH Cl solution: λ, nm (ꢀ, M
-
1
1
1
5
363 s, 1317 m, 1283 w, 1263 m, 1167 m, 1152 m, 1109 m, 1056 m,
015 m, 917 w, 877 w, 860 w, 802 m, 764 m, 734 m, 702 w, 645 w,
36 w, 516 w, 472 w, 434 w, 419 w.
cm )): 686 (1000), 614 (1000), 512 (2000), 450 (4000). Mass
spectrum, ESI + (m/z): 953 M+.
3
0
Cr (dpa) (NCO) (11) was prepared similarly to 10, and cocrys-
3
4
2
7+
Cr
3
(dpa)
4
(NO
3
)
2
(7): In air, 0.10 g (0.11 mmol) of Cr
3
(dpa)
4
Cl
2
tallized with an oxidized Cr
accomplished manually.31
3
species (12). The separation was
and 0.065 g (0.24 mmol) of TlNO
3
were weighed and added to a
Schlenk flask which was then evacuated, purged with nitrogen, and
then charged with 10 mL of acetone while stirring. After 4 h, some
precipitate was visible. The green mixture was filtered over Celite into
a Schlenk tube, and the resulting solution was layered with hexanes.
After 3 days, small crystals were observed on completion of the
diffusion. These were too small for diffraction studies, so the solvents
were removed by decantation and then crystals were dried under
X-ray Crystallography. Crystal data are shown in Table 1. For the
previously reported23 solvates of 1 (1‚benzene, 1‚toluene, 1‚CH Cl ,
2
2
and 1‚THF), the data sets from our archives were re-refined using
SHELXL-9732 without any further data processing. For 1‚benzene and
1‚toluene, the central Cr atoms of each trichromium moiety were
removed and replaced by two atoms each with 0.5 occupancy ∼0.2 Å
in either direction toward the terminal Cr atoms. These atoms were
refined with isotropic thermal parameters, and the positions and relative
occupancies were allowed to refine with the constraint that the sum of
the occupancies for the disordered atoms equal 1. The positions of the
central Cr atoms after convergence of least squares refinement were
different from the starting positions. Thermal ellipsoid plots of 1 from
1‚toluene as obtained in the previously reported refinement and as
vacuum. The green crystals were dissolved in 7 mL of CH
layered with Et O. Block-shaped crystals grew in 2 days. Yield: 23
mg, 21%. Anal. Calcd for Cr : C, 51.07; H, 4.09; N, 18.95.
Found: C, 51.07; H, 4.05; N, 18.45. IR (KBr, cm ): 3448 w, br,
2 2
Cl and
2
3 44 42 14 7
C H N O
-
1
3
1
8
071 w, 3028 w, 2969 w, 2864 w, 2367 w, 2344 w, 1606 s, 1596 s,
548 w, 1468 vs, 1427 vs, 1361 s, 1305 s, 1155 m, 1109 w, 1016 m,
83 w, 859 w, 819 w, 767 m, 741 w, 647 w, 538 w, 519 w, 437 w.
Cr
was weighed and added to a Schlenk flask with 0.098 g (0.34 mmol)
of dried and powdered TlBF . The flask was placed under vacuum and
purged with nitrogen, and 10 mL of CH Cl was added to the flask.
After being stirred for 1.5 h, the mixture was filtered over Celite into
a Schlenk tube and layered with hexanes for crystallization. Diamond-
shaped, dark green crystals were observed within 3 days. Yield: 17
3
(dpa)
4
(BF
4
)
2
(8): In air, 0.152 g (0.17 mmol) of Cr
3
(dpa)
4
Cl
2
(30) Cr
3
(dpa)
4
(NCO)
2
: Anal. Calcd for Cr
3
C
42
H
34
N
14
O
3
: C, 53.73; H, 3.65; N,
-1
2
0.89. Found: C, 53.70; H, 3.57; N, 20.45. IR (KBr, cm ): 3480 w, br,
071 w, 3027 w, 2966 w, 2367 w, 2344 w, 2175 vs, 1605 s, 1594 s, 1547
3
4
w, 1466 vs, 1426 vs, 1364 s, 1311 m, 1280 w, 1155 m, 1108 w, 1053 w,
1
015 m, 917 w, 878 w, 856 w, 763 m, 739 w, 644 w, 627 w, 557 w, 535
w, 517 w, 434 w.
31) Cr (dpa) (NCO)
(NCO)F]PF : IR (KBr, cm ): 3595 w, 3482 w, br, 3074 w, 3030 w,
965 w, 2796 w, 2672 w, 2611 w, 2506 w, 2183 vs, 1751 w, 1604 s, 1550
w, 1465 vs, 1427 vs, 1363 s, 1310 m, 1279 w, 1245 w, 1156 m, 1112 w,
2
2
(
3
4
2 3 4
cocrystallized with the oxidized species [Cr (dpa) -
1
-
6
2
mg, 10%. Anal. Calcd for Cr
3 40 32 12 2 8
C H N B F : C, 47.55; H, 3.19; N,
1
054 w, 1016 w, 919 w, 840 s, 765 m, 740 w, 647 w, 625 w, 557 w, 537
-
1
w, 517 w, 438 w. The axial F anion is probably abstracted from a PF6
1
6.64. Found: C, 47.38; H, 3.22; N, 16.45. IR (KBr, cm ): 3448 w,
anion.
br, 3072 w, 3035 w, 2963 w, 2672 w, 2603 w, 2491 w, 2368 w, 2344
w, 1607 s, 1597 s, 1550 w, 1468 vs, 1427 vs, 1359 s, 1315 m, 1286 w,
(
32) Sheldrick, G. M. SHELXTL97; University of G o¨ ttingen: G o¨ ttingen,
Germany, 1997.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7085

A R T I C L E S
Berry et al.
Table 1. Crystal Data
4
‚2CH
[Cr (dpa)
(NCMe)
(PF ‚2CH
3
CN
1
‚benzene
1‚toluene
Cr (dpa) Cl
1‚CH
Cr (dpa)
CH Cl
2
Cl
2
1‚ether
(dpa) Cl
Et
2‚0.5hexane
Cr (depa) Cl
0.5hexane
3
4
-
5‚CH
Cr (dpa)
CH Cl
2
Cl
2
Cr (dpa) Cl
3
4
2
‚
3
4
2
‚
3
4
Cl
2
‚
Cr
3
4
2
‚
3
4
2
‚
3
2
]
3
4 2
I ‚
compound
formula
C H
6 6
C
7
H
8
2
2
2
O
[Cr
3
(depa)
4
Cl
2
]I
3
6
)
2
3
CN
2
2
C46H38Cl2-
Cr3N12
985.78
C47H40Cl2-
Cr3N12
999.81
C41H34Cl4-
Cr3N12
992.60
C44H42Cl2-
Cr3N12O
981.80
C59H71Cl2-
Cr3N12
1175.18
C56H64Cl2-
Cr3I3N12
1512.79
C48H44Cr3-
F12N16P2
1290.93
C41H34Cl2-
Cr3I2N12
1175.50
FW
crystal
orthorhombic
orthorhombic
orthorhombic
monoclinic
tetragonal
tetragonal
monoclinic
orthorhombic
system
space
group
Pna21
Pca21
Pnn2
P21/c
P 4h n2
I 4h c2
C2/c
Pnn2
a, Å
b, Å
c, Å
â, deg
V, Å
Z
18.634(5)
29.36(2)
16.142(3)
90
8830(6)
8
18.475(2)
14.965(4)
16.469(1)
90
4553(1)
4
12.996(2)
14.1381(8)
11.331(1)
90
2081.9(4)
2
16.1062(9)
15.8183(9)
17.0329(9)
98.471(1)
4292.2(4)
4
19.1659(6)
19.1659(6)
16.080(1)
90
5906.8(5)
4
17.929(1)
17.929(1)
21.828(3)
90
7017(1)
4
28.057(2)
10.1744(7)
22.899(2)
124.674(1)
5376.0(6)
4
13.320(5)
14.299(6)
11.460(5)
90
2182.8(15)
2
3
d(calc),
g cm
Flack
1.483
1.458
1.583
1.519
1.321
1.432
1.595
1.788
-
3
-0.04(2)
-0.03(3)
-0.08(4)
-0.03(2)
c
0.05(2)
parameter
a
b
R1, wR2
0.0540, 0.1203 0.0516, 0.1158 0.0449, 0.1078 0.0522, 0.0913 0.0334, 0.0919 0.0984, 0.2426 0.0320, 0.0872 0.0257, 0.0671
0.0707, 0.1333 0.0646, 0.1251 0.0533, 0.1155 0.0998, 0.1054 0.0388, 0.0992 0.1044, 0.2588 0.0367, 0.0911 0.0282, 0.0692
(
I > 2σI)
a
b
R1, wR2
(all data)
1
2‚3CH
[Cr (dpa)
(NCO)]PF
3CH Cl
2
Cl
2
6
‚CH
(dpa)
CH Cl
2
Cl
2
7‚Et
(dpa)
Et
2
O
(NO
O
8‚2CH
(dpa) (BF
2CH Cl
2
Cl
2
9‚CH
(dpa)
CH Cl
2
Cl
2
10‚2benzene
Cr (dpa) (NCS)
2benzene
10‚2toluene
Cr (dpa) (NCS)
2toluene
11
(dpa)
(NCO)
3
4
F-
Cr
3
4
Br
2
‚
Cr
3
4
3
)
2
‚
Cr
3
4
4
)
2
‚
Cr
3
4
(CN)
2
‚
3
4
2
‚
3
4
2
‚
Cr
3
4
-
6
‚
compound
formula
2
2
2
2
2
2
2
2
2
2
C41H34Br2Cl2-
Cr3N12
1081.52
C44H42Cr3-
N14O7
1034.92
C42H36B2Cl4-
Cr3F8N12
1180.25
C43H34Cl2-
Cr3N14
973.74
C54H44Cr3-
N14S2
1109.15
C56H48Cr3-
N14S2
1137.20
C42H32Cr3-
N14O2
920.82
C44H38Cl6-
Cr3F7N13OP
1297.54
FW
crystal
orthorhombic
monoclinic
monoclinic
orthorhombic
monoclinic
orthorhombic
monoclinic
monoclinic
system
space
group
Pnn2
P21/c
Cc
Pnn2
C2/c
Fdd2
P21/c
P21/c
a, Å
b, Å
c, Å
â, deg
V, Å
Z
13.098(1)
14.203(1)
11.410(1)
90
2122.5(4)
2
16.289(3)
16.374(2)
17.373(1)
103.624(9)
4503.3(9)
4
18.844(3)
17.059(3)
16.439(3)
111.710(3)
4910(1)
4
13.2514(8)
14.1573(8)
11.4203(7)
90
2142.5(2)
2
16.956(6)
18.688(6)
16.949(6)
108.149(6)
5104(3)
4
14.285(1)
24.356(2)
31.508(2)
90
10 962(1)
8
13.869(2)
16.771(3)
17.083(3)
94.928(3)
3959(1)
4
17.310(2)
16.038(2)
19.006(3)
90.782(3)
5276(1)
4
3
d(calc),
g cm
Flack
1.692
1.526
1.597
1.509
1.444
1.378
1.545
1.634
-
3
0.003(8)
c
0.01(2)
0.01(2)
parameter
a
b
R1, wR2
0.0262, 0.0644 0.0818, 0.1684 0.0530, 0.1472 0.0255, 0.0667 0.0377, 0.0977 0.0332, 0.0824 0.0468, 0.1265 0.0521, 0.1312
0.0319, 0.0673 0.1098, 0.1866 0.0550, 0.1497 0.0281, 0.0685 0.0539, 0.1084 0.0387, 0.0864 0.0613, 0.1361 0.0867, 0.1514
(I > 2σI)
a
b
R1, wR2
(all data)
a
R1 ) ∑||Fo| - |Fc||/∑|Fo|. wR2 ) [∑[w(Fo2 - Fc ) ]/∑[w(Fo ) ]] , w ) 1/σ (Fo ) + (aP) + bP, where P ) [max(0 or Fo ) + 2(Fc )]/3. Refined
b
2 2
2 2 1/2
2
2
2
2
2
c
as a racemic twin.
obtained by the procedure just described are shown in parts a and b,
respectively, of Figure 1. It should be noted that the elongated thermal
ellipsoid for the central Cr atom in Figure 1a indicates the presence of
this unresolved disorder. The difference between the long and short
Cr-Cr distances, ∆dCr-Cr, is ∼0.23 Å. This difference is statistically
real because, even in the very worst case, it is >20 times the standard
deviation. The occupancies of the central Cr atom positions converged
to the following ratios: 65:35 and 57:43 for the two independent
molecules in 1‚benzene, and 52:48 for 1‚toluene.
half atoms converged to positions about 0.1 Å off of the two-fold axis.
Here, the disordered central Cr atom is required to have an occupancy
of 0.5 by the crystallographic 2-fold symmetry.
By re-refining the structure of 1‚THF similarly, it was found that
the central Cr atom did move back to its initial position of 222
symmetry. Attempts were made to crystallize 1 from THF again to get
a data set of higher quality, but there were difficulties as discussed
later.
For 1‚Et
2
O, 2‚0.5hexane, 3, 4‚2CH
3
CN, 5‚CH
2
Cl
2
, 6‚CH
2
Cl
2
,
2
,
In the previously reported structure of 1‚CH
distances were constrained to be equal by the crystallographic two-
fold axis that passes through the midpoint of the Cr chain. In our re-
2
Cl
2
, the two Cr-Cr
8‚2CH Cl , 9‚CH
2
2
2
Cl , 10‚2benzene, 10‚2toluene, 11, and 12‚3CH
2
2
Cl
for which structures have not previously been reported, suitable crystals
were mounted either on the end of a quartz fiber or on a nylon loop
and transferred to the goniometer of a Bruker SMART CCD area
detector diffractometer and cooled to -60 °C. Geometric and intensity
3
refinement of this structure, however, the central Cr atom was split
into two and each half moved ∼0.2 Å off of the two-fold axis toward
a terminal Cr atom. After several cycles of least squares refinement
with an occupancy of 0.5 and an isotropic thermal parameter, the central
33
data were collected using SMART software and were processed using
SAINT software34 into SHELX format. Corrections for absorption were
7086 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
A R T I C L E S
Figure 3. Thermal ellipsoid plot of 3 viewed along the Cr3 axis with
ellipsoids drawn at the 30% probability level. Hydrogen atoms are removed,
and only one orientation of the disordered ethyl groups is shown.
Figure 1. Thermal ellipsoid plot of 1‚toluene (a) as it appears in ref 23
and (b) as reinterpreted here with ellipsoids drawn at the 50% probability
level. Solvent molecules and hydrogen atoms have been omitted.
Figure 4. Thermal ellipsoid plot of 4 from 4‚2CH3CN with ellipsoids drawn
at 50% probability. Pyridyl hydrogen atoms and one of the two orientations
of the disordered central chromium atom have been omitted for clarity.
In most cases, the space group was determined unambiguously
from the systematic absences in the data. For those which were
not determined unambiguously, the choice of centrosymmetric
versus noncentrosymmetric groups was made on the basis of the
intensity statistics of the data and successful structure refinements. The
structures were solved either by direct methods or by the Patterson
method and refined using SHELXL-97.3 Crystal data are listed in
Table 1, and important interatomic distances are shown in Table 2.
For all structures in noncentrosymmetric space groups, the correct
choices of polar axes were assured by refinement of the Flack
parameter.40 Compounds 3 and 8 were refined as racemic twins. Figures
2,39
2
-6 show thermal ellipsoid plots of compounds 2, 3, 4, 9, and 10,
Figure 2. Thermal ellipsoid plot of 2 with ellipsoids drawn at the 30%
probability level. Hydrogen atoms have been removed, and only one
orientation of the disordered ethyl groups is shown.
respectively.
(
36) Pflugrath, J.; Messerschmidt, A. MADNES, Munich Area Detector (New
EEC) System, Version EEC 11/1/89, with enhancements by Enraf-Nonius
Corp., Delft, The Netherlands. A description of MADNES appears in:
Messerschmidt, A.; Pflugrath, J. J. Appl. Crystallogr. 1987, 20, 306.
37) (a) Kabsch, W. J. Appl. Crystallogr. 1988, 21, 67. (b) Kabsch, W. J. Appl.
Crystallogr. 1988, 21, 916.
applied using SADABS.35 A crystal of 6‚Et
2
O was mounted on the tip
of a quartz fiber and transferred to the goniometer of a Nonius FAST
(
area detector system. Geometric and intensity data were collected using
3
6
(38) Program for absorption correction for Enraf-Nonius FAST diffractometer
using the method of: Blessing, R. H. Acta Crystallogr. 1995, A51, 33.
(39) The refinement of most of the crystal structures presented was routine except
with respect to the disordered metal atom positions. In 2‚0.5hexane and 3,
several of the ethyl groups of the depa ligands were found to be disordered
in two positions whose relative occupancies were refined. Many of the
structures containing disordered interstitial solvent molecules were refined
the software program MADNES and processed into SHELX format
37
by the program PROCOR. The program SORTAV was used to correct
3
8
for absorption.
(
(
(
33) SMART V5.618 Software for the CCD Detector System; Bruker Analytical
X-ray Systems, Inc.: Madison, WI, 1998.
with distance constraints on the disordered moieties. In 7‚Et
2 2
O and 8‚2CH -
34) SAINTPLUS, V6.45A Software for the CCD Detector System; Bruker
Analytical X-ray System, Inc.: Madison, WI, 1998.
Cl , in which all of the Cr atoms and the axial ligands are disordered, the
2
two orientations of the molecule were restrained to have a similar geometry,
although it should be pointed out that in the case of 8, the geometry of the
molecule is slightly different in the major and minor orientations.
(40) Flack, H. D. Acta Crystallogr. 1983, A39, 876.
35) SADABS V2.10. Program for absorption correction using SMART CCD
data based on the method of Blessing: Blessing, R. H. Acta Crystallogr.
1
995, A51, 33.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7087

A R T I C L E S
Berry et al.
Table 2. Selected Interatomic Distances for Compounds 1-12
compound
Cr−Cr, Å
Cr‚‚‚Cr, Å
Cr−N (inner), Å
Cr−N (outer), Å
Cr−L (axial), Å
2.532[2]
∆dCr-Cr, Å
Cr1‚‚‚Cr3, Å
1
‚benzene
2.227[9]
.236[9]
2.24[1]
2.483[9]
2.481[9]
2.48[1]
2.029[6]
2.028[6]
2.031[7]
2.027[6]
2.031[4]
2.022[3]
2.03[1]
2.036[4]
2.014[4]
2.026[3]
2.036[6]
2.118[5]
2.116[5]
2.123[6]
2.114[5]
2.115[3]
2.134[3]
2.101[9]
2.101[1]
2.114[3]
2.122[2]
2.107[7]
0.256[9]
0.245[9]
0.24[1]
0.223(4)
0.220[4]
0
0.295(8)
0.321(3)
0.298(1)
0.259(1)
0.712(5)
4.710[9]
4.719[9]
4.718(2)
4.731(2)
4.717[4]
4.7560(5)
4.587(8)
4.607(3)
4.666(1)
4.705(1)
4.578(5)
2
2.544[2]
2.556[2]
2.550(2)
2.530[1]
2.5122(9)
2.356(4)
2.339(2)
2.996(1)
2.7364(4)
2.09(1)
1
1
1
2
3
4
5
6
7
‚toluene
‚CH2Cl2
‚Et2O
2.254(4)
2.249[4]
2.3780(5)
2.146(8)
2.143(3)
2.184(1)
2.223(1)
1.934(5)
2.477(4)
2.469[4]
symmetrical
2.441(8)
2.464(3)
2.482(1)
2.482(1)
2.644(5)
‚0.5hexane
‚2CH3CN
‚CH2Cl2
‚CH2Cl2
a
‚Et2O
2
.30(1)
2.204(7)
.347(6)
8
‚2CH2Cl2a
1.963(5)
2.590(5)
2.034[9]
2.106[6]
0.627(5)
4.553(5)
2
9
1
1
1
1
‚CH2Cl2
0‚2benzene
0‚2toluene
1
2.3700(3)
2.234(1)
2.215(1)
2.26[1]
symmetrical
2.482(1)
2.465(1)
2.48[1]
2.032[2]
2.039[2]
2.020[3]
2.030[4]
2.014[3]
2.119[2]
2.113[2]
2.116[2]
2.118[3]
2.284(2)
2.203(2)
2.203(3)
2.177[4]
2.134(4) (NCO)
1.868(3) (F)
0
4.7400(3)
4.716(1)
4.680(1)
4.735(1)
4.566(1)
0.248(1)
0.249(1)
0.22[1]
0.586(1)
2‚3CH2Cl2
1.9905(9)
2.576(1)
2.106[3] (NCO)
2.075[4] (F)
a
Bond distances from only the major orientation of 7 and 8 are given. The corresponding distances for the minor orientation have larger esd’s and may
not be accurate.
reaction mixture, reaction of Lidepa and CrCl2 in a 4:3 ratio in
refluxing THF generates Cr3(depa)4Cl2 (2) in solution. As
41
27
already noted for the analogous cobalt and nickel compounds,
complexes with the ethyl-substituted depa ligand are much more
soluble than those of the dpa ligand.
THF
3
CrCl + 4Lidepa
8 Cr (depa) Cl + 4LiCl (1)
3 4 2
∆
2
Oxidation of 2 with the easily prepared and mild oxidant
7
+
ferrocenium triiodide yields the Cr3 species 3, similarly to
3+
the previously described syntheses of related Cr3(dpa)4
species.
4
2
For the syntheses of Cr3(dpa)4X2 complexes with X other
than Cl, two methods are useful. The first is simple metathesis
with thallium reagents (eq 2).
Figure 5. Thermal ellipsoid plot of 9 with ellipsoids drawn at 50%
probability. Hydrogen atoms and interstitial solvents have been omitted.
CH Cl
2
2
Cr (dpa) Cl + 2TlX
8 Cr (dpa) X + 2TlCl (2)
3 4 2
3
4
2
Silver reagents are not used to avoid oxidation. For example,
while reaction of 1 with 1 equiv of AgBF4 yields the unsym-
metrical compound Cr3(dpa)4Cl(BF4), 2 equiv of AgBF4 is
7
+ 19
enough to oxidize the core to Cr3
.
Thus, the nonoxidizing
TlBF4 salt must be used to prepare Cr3(dpa)4(BF4)2, 8. An
analogous reaction with TlNO3 yields Cr3(dpa)4(NO3)2, 7.
Unfortunately, only low yields of 7 and 8 were obtained. This
6+
difficulty prompted us to seek another route to Cr3 complexes
with labile axial ligands.
It has been found that reaction of 1 with TlPF6 in aceto-
nitrile produces [Cr3(dpa)4(NCMe)2](PF6)2, 4 (eq 3), in high
yield, and this is a useful starting material for the synthesis
of other Cr3(dpa)4X2 complexes by addition of X , as shown
in eq 4. Most Cr3(dpa)4X2 complexes precipitated readily from
the MeOH, and those that did not could be extracted from
Figure 6. Thermal ellipsoid plot of 10 with ellipsoids drawn at 30%
probability. Hydrogen atoms were removed.
-
3. Results and Discussion
Syntheses. The syntheses of trichromium compounds with
the dpa ligand and with the ethyl-substituted homologue (depa)
proceeded similarly. Whereas Cr3(dpa)4Cl2 precipitates from the
(41) Berry, J. F.; Cotton, F. A.; Lu, T.; Murillo, C. A. Inorg. Chem. 2003, 42,
4425.
(
42) Cl e´ rac, R.; Cotton, F. A.; Daniels, L. M.; Dunbar, K. R.; Murillo, C. A.;
Pascual, I. Inorg. Chem. 2000, 39, 752.
7088 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
the residue once the MeOH was removed.
MeCN
A R T I C L E S
Scheme 6
Cr (dpa) Cl + 2TlPF
8
3
4
2
6
[
Cr (dpa) (NCMe) ](PF ) + 2TlCl (3)
3 4 2 6 2
MeOH
-
[
Cr (dpa) (NCMe) ](PF ) + 2X
8
3
4
2
6 2
-
Cr (dpa) X + 2PF (4)
3
4
2
6
It should be noted that reaction of Cr3(dpa)4Cl2 with excess
NaCN in methanol does not produce Cr3(dpa)4(CN)2, but instead
the bright red dinuclear compound Cr2(dpa)4 precipitates from
the reaction mixture in quantitative yield. This quadruply bonded
dichromium complex has been described before.43 Although we
did not persue the isolation of the other products of this reaction,
these are probably Na4Cr(CN)6 and NaCl. Because of this, it is
necessary to control the stoichiometry carefully and to use 4 in
reactions with a stoichiometric amount of NaX as described in
eq 4. Compound 4 in solution is very air sensitive, and solutions
rapidly turn from green to orange upon exposure to the
atmosphere. The reactions described by eq 4 must be carried
thermal displacement ellipsoids representing the atoms will be
small and nearly spherical. It is especially important to pay
attention to the appearance of the central atom.
If the trimetal chain has one shorter and one longer distance,
one might na ¨ı vely expect to see a set of ellipsoids as shown in
(b) or (c). However, a problem arises because the intermolecular
forces that determine the packing of such molecules are short-
range and depend on how the outermost atoms of each molecule
interact with the outermost atoms of their neighbors. As a result,
such molecules tend to occupy the same loci within the limits
of detection, and thus their thermal displacement ellipsoids are
small and spherical. As for the central atom, it will be an atom
having fractional occupancy at each of two sites close to the
center (d). Because the vibrational amplitude along the chain
direction is equal to or greater than the separation between the
loci of the fractional atoms, the normal anisotropic structural
refinement procedure will result in the appearance of only one
full atom at the center. However, the clue that this is not really
the case will be an abnormally elongated shape (prolate
spheroid) of the thermal vibration ellipsoid, as shown in (e).
When this is seen, it is necessary to carry out a refinement of
two partial atoms with coordinates and occupancies as param-
eters to get a true picture of the unsymmetrical three-atom chain.
The importance of the difference between the two sets of
displacement ellipsoids (a) and (e) cannot be overemphasized.
If an error is made in this stage of interpreting the structures
that is, taking (e) to be equivalent to (a)sinstead of realizing
that the structure is a disordered superposition of (b) and (c),
there is no hope of finally achieving an understanding of the
bonding in the molecule. Here, we provide definitive clarifica-
tion of this question with regard to compounds 1-12.
out under nitrogen to avoid oxidation to Cr37 species. As
+
31
with the previously reported analogues of nickel, [Ni3(dpa)4-
NCMe)2](PF6)2,2
7,44
45
(
4
(
and cobalt, [Co3(dpa)4(NCCH3)2](PF6)2,
is a valuable starting material, useful for the synthesis of Cr3-
dpa)4X2 complexes with various X ligands. While replacement
of the axial ligands in tricobalt units is also easily carried out
4
6
by use of the tetrafluoroborate complex Co3(dpa)4(BF4)2, the
analogous trichromium tetrafluoroborate complex is much more
air sensitive and has been synthesized only in low yields.
Recently, the preparation of Cr3(dpa)4(NCS)2 (10) was
reported by another group by reaction of CrCl2, Hdpa, and
t
KO Bu, in molten naphthalene followed by stirring with KSCN
in a solution open to air for 2 days.47 The yield was not given.
Furthermore, the reported crystal structure of 10 is not in
agreement with the two we report here (vide infra), and the
compound was said to be red. Compound 10 as prepared by
our method is analytically pure and is dark green. On the basis
of the reaction conditions described by the other group and the
discussion above, it is likely that the reported method for the
synthesis of Cr3(dpa)4(NCS)2 yields a mixture of 10 and the
bright red Cr2(dpa)4. Thus, to prepare 10 in yields as high as
4
8
49,50
The compound Cr3(dpa)4Cl2, like its cobalt, nickel,
copper
and
5
1,52
analogues, crystallizes in many solvates depending
8
8% and in good purity, our method involving 4 should be used.
on the solvent used for crystallization. We report here a new
Crystal Structures. Although trichromium EMACs have
crystal form of Cr3(dpa)4Cl2, 1‚Et2O, which prompted us to
only recently been discovered, they already pose both experi-
mental and theoretical challenges. As always in science,
theoretical efforts cannot succeed without qualitatively and
quantitatively reliable data. In dealing with the crystal structures
of trichromium EMACs, a particular type of crystallographic
disorder often poses an insidious problem. This is represented
in Scheme 6. If the trimetal chain is really symmetric (a), the
23
reinvestigate the four previously reported structures of 1: 1‚
benzene, 1‚toluene, 1‚CH2Cl2, and 1‚THF. The reason for this
reinvestigation is that, in 1‚Et2O, anisotropic refinement of all
three Cr atoms revealed that the displacement ellipsoid of the
central Cr atom was slightly elongated in the direction of the
Cr-Cr vectors as in (e). It could be better refined after being
1
split into two atoms with ∼ /2 occupancy. Thus, the molecule
(
43) (a) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Pascual, I.; Zhou, H.-C.
J. Am. Chem. Soc. 1999, 121, 6856. (b) Edema, J. J. H.; Gambarotta, S.;
Meetsma, A.; Spek, A. L.; Smeets, W. J. J.; Chiang, M. Y. J. Chem. Soc.,
Dalton Trans. 1993, 789.
(48) Cl e´ rac, R.; Cotton, F. A.; Daniels, L. M.; Dunbar, K. R.; Murillo, C. A.;
Wang, X. Inorg. Chem. 2001, 40, 1256.
(49) Berry, J. F.; Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Wang, X. Inorg.
Chem. 2003, 42, 2418.
(
44) Berry, J. F.; Cotton, F. A.; Murillo, C. A. Dalton Trans. 2003, 3015.
45) Cl e´ rac, R.; Cotton, F. A.; Dunbar, K. R.; Lu, T.; Murillo, C. A.; Wang, X.
Inorg. Chem. 2000, 39, 3065.
(50) Aduldecha, S.; Hathaway, B. J. Chem. Soc., Dalton Trans. 1991, 993.
(51) Berry, J. F.; Cotton, F. A.; Lei, P.; Murillo, C. A. Inorg. Chem. 2003, 42,
377.
(52) (a) Wu, L.-P.; Field, P.; Morrissey, T.; Murphy, C.; Nagle, P.; Hathaway,
B.; Simmons, C.; Thornton, P. J. Chem. Soc., Dalton Trans. 1990, 3853.
(b) Pyrka, G. J.; El-Mekki, M.; Pinkerton, A. A. J. Chem. Soc., Chem.
Commun. 1991, 84.
(
(46) Cl e´ rac, R.; Cotton, F. A.; Jeffery, S. P.; Murillo, C. A.; Wang, X. Inorg.
Chem. 2001, 40, 1265.
(
47) Lin, S.-Y.; Chen, I.-W. P.; Chen, C.-h.; Hsieh, M.-H.; Yeh, C.-Y.; Lin,
T.-W.; Chen, Y.-H.; Peng, S.-M. J. Phys. Chem. B 2004, 108, 959.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7089

A R T I C L E S
Berry et al.
Table 3. Comparison of Previously Reported and Newly
Corrected Interatomic Distances for 1
symmetry, in which the two Cr-Cr distances of 2.3780(5) Å
are crystallographically equivalent. The thermal ellipsoid for
the central Cr atom in 2‚0.5hexane is not elongated, which
strongly suggests that a truly symmetrical molecular structure
and not a disordered unsymmetrical structure is present.
Although this differs from the results of the re-refined structures
previously reported
Cr−Cr distances, Åa
compound
corrected Cr−Cr, Åb
corrected Cr‚‚‚Cr, Åb
1
‚benzene 2.296(2), 2.414(2)
2.227[9]
2.483[9]c
2.481[9]
2.48[1]
2e.477(4)
c
c
2
.326(2), 2.390(2)
2.236[9]
1
1
1
1
‚toluene
‚CH2Cl2
‚THF
2.353(2), 2.365(2)
2.24[1]
d
6+
2.366(1)
2.254(4)
of 1, it is credible, since we have recently reported Cr3 chains
d
2.365(2)
not resolved
58,59
of ligands other than dpa
and found that, in some cases,
‚Et2O
2.249[4]
2.469[4]
certain equatorial ligands (such as unsymmetrical formamidi-
6
+
58
a
Data from ref 23. b The esd’s of these measurements are rather high
nates) favor symmetrical Cr3 chains, as will be discussed
in more detail later.
c
due to the disorder of the metal atom positions. This structure has two
molecules in the asymmetric unit. The two Cr-Cr distances are related
by symmetry. See ref 57.
d
6+
The more basic depa ligand also allows the Cr3 chain to
e
7
+
be more easily oxidized to Cr3 (vide infra). The structure of
the oxidized species 3 is shown in Figure 3. Although the Cr3
chain lies on a position of 222 symmetry, it is possible to sepa-
rate the central Cr atom into two and resolve the asymmetry of
has one short Cr-Cr distance of 2.249[4] Å, and a longer
distance, 2.469[4] Å, to the third Cr2+ ion.
Reinvestigation of the structures of 1‚benzene, 1‚toluene, and
‚CH2Cl2 by re-refinement of the data sets from our archives
7+
the compound. The Cr3 chain is distinctly unsymmetrical with
1
a short Cr-Cr distance of 2.146(8) Å and a longer distance of
showed that these three structures, which were previously
reported to be of more or less symmetrical molecules, are also
better described as unsymmetrical, as shown in Figure 1. The
unusually elongated thermal ellipsoid for the central Cr atom
3
+
2
.441(8) Å to the isolated Cr ion, as seen for the previously
7
+
42
reported Cr3 compounds. It is worth noting that the overall
length of the Cr3 chain (i.e., the distance from Cr1 to Cr3) short-
ens from 4.7650(5) to 4.587(8) Å upon oxidation. Because the
central Cr atom is disordered in two positions, the terminal
Cr-N distances do not reflect the differences expected for
Cr(II)-N versus Cr(III)-N bond distances, but have an aver-
(Figure 1a) has its major axis along the Cr-Cr vector, and this
is the best indicator of this type of disorder, which we have
7+
42
noted before in the crystal structures of Cr3 compounds and
1
0+
53
10+
Cr5
compounds. It should be mentioned that another Cr5
-
54
aged value of 2.101[9] Å. The I3 counteranion is disordered
compound, Cr5(tpda)4(NCS)2 (tpda ) tripyridyldiamide), and
5
5
over a site of 4h symmetry and also has unequal I-I distances
a heptanuclear chromium compound, Cr7(teptra)4Cl2 (teptra
tetrapyridyltriamide), were reported to have almost evenly
7
+
of 2.848(7) and 2.98(1) Å. In short, both the Cr3 chain and
)
-
the I3 chain in this compound are unsymmetrical in the solid
spaced Cr-Cr distances even though some of the chromium
atoms showed distinctly elongated ellipsoids. We have recently
resynthesized the former and shown that the compound is
unsymmetrical with alternating long-short-long-short Cr-
Cr distances. In 1‚THF, the molecule crystallizes on a position
of 222 symmetry, and attempts to separate the central Cr atom
into two components were made, but these were not successful.
However, in a reexamination of newly grown crystals of 1‚
THF, we have observed a complex diffraction pattern, indicative
of twinning. After many unsuccessful attempts at growing better
crystals, we do not believe it is possible to obtain a quality
state.
Compounds 5 and 6, like 1, crystallize from dichloromethane
as monodichloromethane solvates in the noncentric space
group Pnn2, and again the central Cr atoms in 5‚CH2Cl2 and
56
6
‚CH2Cl2 may be split into two entities, thus resulting in
unsymmetrical molecules with short Cr-Cr distances of 2.184(1)
and 2.223(1) Å, and long Cr‚‚‚Cr separations of 2.482(1) and
2
.482(1) Å for 5 and 6, respectively. The crystal structures of
4
, 10, and 11 also contain unsymmetrical molecules which are
4
7
disordered similarly. In a recently reported crystal structure
of 10, this disorder was not taken into account. The central Cr
atom was shown with an elongated thermal ellipsoid and Cr-
Cr distances of 2.277(2) and 2.391(2) Å. These do not agree
with the ones found here in two independent structures of 10
in which the short Cr-Cr distances are 2.234(1) and 2.215(1)
Å and the long Cr‚‚‚Cr separations are 2.482(1) and 2.465(1)
Å for 10‚2benzene and 10‚2toluene, respectively. The disorder
of the central Cr atom positions must be taken into account for
the proper Cr-Cr bond lengths to be revealed.
structure of 1‚THF, and it will be omitted from further
discussions.57 A comparison of the previously reported and new
interatomic distances for 1 is given in Table 3.
The ethyl-substituted Cr3(depa)4Cl2, 2, shown in Figure 2,
crystallizes in the space group P 4h n2 on a position with 2-fold
(
53) (a) Cotton, F. A.; Daniels, L. M.; Lu, T.; Murillo, C. A.; Wang, X. J. Chem.
Soc., Dalton Trans. 1999, 517. (b) Cotton, F. A.; Daniels, L. M.; Murillo,
C. A.; Wang, X.; Murillo, C. A. Chem. Commun. 1999, 2461.
54) Chang, H.-C.; Li, J.-T.; Wang, C.-C.; Lin, T.-W.; Lee, H.-C.; Lee, G.-H.;
Peng, S.-M. Eur. J. Inorg. Chem. 1999, 1243.
(
(
(
(
Compounds 7 and 8 presented especially severe crystal-
lographic disorder. All three metal atoms in 7‚Et2O and
55) Chen, Y.-H.; Lee, C.-C.; Wang, C.-C.; Lee, G.-H.; Lai, S.-Y.; Li, F.-Y.;
Mou, C.-Y.; Peng, S.-M. Chem. Commun. 1999, 1667.
56) Berry, J. F.; Cotton, F. A.; Fewox, C.; Lu, T.; Murillo, C. A.; Wang, X.,
manuscript in preparation.
8
‚2CH2Cl2 displayed elongated thermal ellipsoids which sug-
57) Using only the most intense reflections, it was possible to index the pattern
to the previously reported tetragonal cell. Yet this does not account for all
of the diffraction peaks. More careful inspection revealed that the pattern
gested that the three “chromium atoms” are really pairs of partial
Cr atoms in close proximity. There are two possible models
for overlapping Cr3 molecules, as shown in Scheme 7, depend-
ing on how the central Cr atoms are assigned.
could be indexed to an orthorhombic cell isomorphous to that in 1‚CH
Cl , which is twinned in four components, all related to one another by
0° rotation about one of the unit cell axes. We believe that the
2
-
2
9
orthorhombic cell is correct and merely simulates the tetragonal symmetry
by this complex twinning. Unfortunately, satisfactory refinement of the
structure in this Pnn2 setting was not possible (probably due to incorrect
intensity data arising from overlapping reflections from the four twin
components), although the THF molecule here is ordered on a position of
(58) (a) Cotton, F. A.; Lei, P.; Murillo, C. A.; Wang, L.-S. Inorg. Chim. Acta
2003, 349, 165. (b) Cotton, F. A.; Lei, P.; Murillo, C. A. Inorg. Chim.
Acta 2003, 349, 173.
(59) (a) Cl e´ rac, R.; Cotton, F. A.; Daniels, L. M.; Dunbar, K. R.; Murillo, C.
A.; Zhou, H.-C. Inorg. Chem. 2000, 39, 3414. (b) Cotton, F. A.; Daniels,
L. M.; Lei, P.; Murillo, C. A.; Wang, X. Inorg. Chem. 2001, 40, 2778. (c)
Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Wang, X. Chem. Commun.
1998, 39.
2
-fold symmetry (disordered over a 222 position in P 4h n2), and the central
Cr atom could be refined off the two-fold axis resulting in an unsymmetrical
structure. While it is not clear that this twinning necessarily occurred in
the previous work on the structure of 1‚THF, we believe this to be so.
7090 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
A R T I C L E S
Scheme 7
has completely different physical properties, and the struc-
ture of the molecule changes dramatically depending on the
temperature at which the X-ray experiment is done. For
simplicity, only the structures of s-Co3(dpa)4Cl2‚CH2Cl2 and
u-Co3(dpa)4Cl2‚2CH2Cl2 at 298 K are given in Table 4. Several
generalizations can be made about these data. First, unsym-
6
+
metrical Cr3 complexes have the shortest M-M distances of
all the compounds presented, and this can be attributed to the
4
+
formation of a localized Cr2
quadruple bond. Tricobalt
complexes are the only other trimetal species with unsym-
metrical structures, but of all the symmetrical complexes,
tricobalt species have the shortest M-M distances, and one may
6
+
suppose, perhaps na ¨ı vely, that symmetrical Co3 chains have
a higher bond order. Symmetrical trichromium complexes have
Cr-Cr distances of intermediate length. The Ni3 complexes,
which are all symmetric, have the longest interatomic distances,
6
+
since there are no Ni-Ni bonds in Ni3 EMACs.
6
+
The most notable difference between the Cr3 complexes
6
+
6+
and their Co3 and Ni3 analogues, however, is that the Cr-
Table 4. Comparison of M3(dpa)4X2 Species (M ) Cr, Co, Ni)
Lax distances are uniformly longer by as much as 0.29 Å in the
Cr3 complexes. This difference is large as compared to the
compound
M−M, Å
M‚‚‚M, Å
M−Lax
ref
6+
1
s-Co
u-Co
Ni
2.241[7]a
2.3369(4)
2.299(1)
2.478[7]a 2.542[2]a
this work
difference in the respective atomic radii of Cr, Co, and Ni which
b
a
b
3
(dpa)
(dpa)
(dpa) Cl
4
Cl
2
)
)
2.4881(8) 22
60
differ only by ∼0.03 Å. It might therefore appear that the
b
2.471(1)b 2.399(2)
b
3
4
Cl
2
22
49
this work
45
27
this work
46
44
this work
46
a
axial ligands are not bound very tightly, perhaps because Cr-L
antibonding orbitals are significantly populated. Despite the long
bond lengths to the axial ligands, many of these complexes show
the presence of a molecular ion in their mass spectra. Moreover,
3
4
2
2.4313[9]
2.143(3)
2.333[2]
4
2.464(3) 2.339(2)
a
a
[
[
Co
Ni
3
(dpa)
(dpa)
4
(NCCH
(NCCH
3
)
2
](PF
](PF
6
)
2
2.301[1]
)
2.089[5]
2.049[7]
2.284(2)
2.040(2)
2.014(2)
3
4
3
)
2
6
)
2
2.374[2]
)
)
)
)
9
Co
Ni
1
Co
Ni
1
Co
Ni
2.3700(3)
2.3392(2)
2.4523(3)
6
+
3
(dpa)
(dpa)
4
(CN)
(CN)
2
Co3 complexes have 21 electrons with which to fill the same
molecular orbitals, and there is no elongation of the Co-L
distances. Thus, at this time, there is no unambiguous explana-
tion for the long Cr-L distances. It may be noted, however,
that a similar elongation occurs in dichromium quadruply
bonded species, as shown by the long Cr-Lax distances (in the
range from 2.21 to 2.74 Å for various axial ligands) observed
3
4
2
a
2.473[1]a 2.203[3]
a
a
0
2.224[1]
2.313[1]
a
3
(dpa)
(dpa)
4
(NCS)
(NCS)
2
)
)
)
2.050[5]
1.988(6)
2.24[1]
3
4
2
2.4285(9)
2.430[2]
2.344(1)
47
28
28
28, 44
3
3
(dpa)
(dpa)
4
(CCPh)
(CCPh)
2
2.401(1) 2.022[5]
a
a
3
4
2
2.472[1]
)
2.014[8]
a
Values are averaged from the different crystal structures reported for
11
in Cr2(carboxylate)4(Lax)2 complexes. Also, the Cr-NCCH3
b
this compound. Values are from the 298 K crystal structures.
distances of 2.339(2) Å in 4 are almost equal to those in
Cr2(carboxylate)4(NCCH3)2 complexes (2.326[5] Å).14 This is
in marked contrast to the Cr-NCCH3 distances typically
observed for mononuclear Cr(II) species where no Cr-Cr bond
These two models (and hence the correct Cr-Cr bond
distances) would be indistinguishable if the occupancies of the
two components were each 50%. Fortunately, for 7‚Et2O and
6
1
6+
is present (∼2.07 Å). Since the unsymmetrical Cr3 com-
8
‚2CH2Cl2, the occupancy ratios are about 70:30. Thus, one
4
+
plexes can be described as having an isolated Cr2 quadruple
bond, it is reasonable that di- and trinuclear species have similar
Cr-Lax bond distances.
set of three 70% occupied chromium sites and one set of
three 30% occupied sites may be chosen, thus allowing for
the satisfactory assignment of model 2 for both 7‚Et2O and
It is also useful to compare these trichromium complexes with
8
‚2CH2Cl2. Thus, each of these compounds is very unsym-
6+
related Ru2 compounds having various axial ligands reported
metrical with ∆dCr-Cr ≈ 0.7 Å, and each has a Cr-Cr quadruple
bond distance shorter than 2 Å. It should be noted that the axial
ligands in 7 and 8 also display positional disorder since the
6
2-64
65
by Ren
and Bear. For these Ru2(L)4X2 compounds, there
is a dramatic difference depending on whether X is a strong
donor ligand (CN or CCPh) or a weaker one (e.g., I or
II
Cr -Lax distance from the isolated Cr atom (2.10-2.20 Å) is
6
4
NCNCN). The former complexes are diamagnetic and have
significantly shorter, as would be expected, than the Cr‚‚‚Lax
distance from the quadruply bonded Cr atom (>2.30 Å).
Crystals of 9‚CH2Cl2 are isostructural to 1‚CH2Cl2, 5‚
CH2Cl2, and 6‚CH2Cl2 and crystallize in the noncentrosymmetric
space group Pnn2. Yet compound 9 is crucially different from
the others as it contains symmetrical molecules, since the
ellipsoid for the central Cr atom is nearly spherical.
long Ru-Ru distances, whereas the latter are paramagnetic with
6
4
short Ru-Ru bonds. It is suggested that in Ru2(L)4X2
complexes with Ru-C bonds (i.e., with X ) CN or CCPh),
the Ru2 σ orbitals are used mainly for the formation of the Ru-C
(
(
(
60) International Tables for Crystallography; Wilson, A. J. C., Ed.; Kluwer
Academic Publishers: Dordrecht, 1995; Vol. C, p 681.
61) Henriques, R. T.; Herdtweck, E.; K u¨ hn, F. E.; Lopes, A. D.; Mink, J.;
Rom a˜ o, C. C. J. Chem. Soc., Dalton Trans. 1998, 1293.
62) Lin, C.; Ren, T.; Valente, E. J.; Zubkowski, J. D. J. Chem. Soc., Dalton
Trans. 1998, 571.
It is useful to compare the structures of the Cr3(dpa)4X2
complexes to their Co3(dpa)4X2 or Ni3(dpa)4X2 analogues. For
convenience, the relevant data have been collected in Table 4,
and, where several structures are known, these data have been
averaged except for Co3(dpa)4Cl2. Each Co3(dpa)4Cl2 solvate
(63) Xu, G.; Campana, C.; Ren, T. Inorg. Chem. 2002, 41, 3521.
(
(
64) Chen, W.-Z.; Ren, T. Inorg Chem. 2003, 42, 8847.
65) Bear, J. L.; Han, B.; Huang, S.; Kadish, K. M. Inorg. Chem. 1996, 35,
3012.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7091

A R T I C L E S
Berry et al.
and the value of θ for 13 is larger by an order of magnitude.
This difference between the magnetic properties of symmetrical
and unsymmetrical compounds may originate in the zero-field
splitting which is probably the most important contributing factor
to the magnitude of θ. In a symmetrical complex with four
unpaired electrons delocalized over three metal atoms, the spin-
orbit coupling and hence the zero-field splitting is expected to
be larger than for a compound with four unpaired electrons
localized on a single metal atom. This corresponds to a more
pronounced decrease of øT at lower temperatures for the
symmetrical complexes.
It might be questioned whether the symmetrical complexes
with Cr-Cr distances over 2.35 Å contain metal-metal bonds
at all. For instance, the Cr-Cr distances in 13 of 2.427[2] Å
are close to those Ni‚‚‚Ni separations in Ni3(dpa)4Cl2 (2.4313[9]
4
9
Å) which does not contain Ni-Ni bonds. The difference
between 13 and Ni3(dpa)4Cl2 lies in their electronic structures.
If a delocalized electronic structure were to be invoked for the
latter, its 24 d electrons would fill and occupy all of the bonding,
nonbonding, and antibonding orbitals leaving no net bond.
Experimentally, though, Ni3(dpa)4Cl2 is best described as having
Figure 7. Plots of øT (green circles) and 1/ø (blue squares) versus T for
7
. The solid line along the blue data represents a fit of the data to the Curie-
Weiss expression for S ) 2. The refined parameters are given in the lower
right-hand portion of the plot.
Table 5. Magnetic Susceptibility Data for 1-4, 7, 9, 10, and 13
compound
g (Land e´ )
θ (Weiss constant, K)
ref
2
+
a diamagnetic square-planar central Ni ion flanked by two
1
2
3
7
9
2.08
NRa
23
2+
terminal high-spin S ) 1 Ni ions which couple antiferromag-
1.9845(6)
2.0100(1)
1.9774(1)
1.9798(4)
2.0443(3)
1.924(2)
-0.87(6)
-0.26(1)
-0.23(1)
-0.83(4)
-0.13(3)
-3.2(2)
this work
this work
this work
this work
this work
28
4
9
netically with each other. If there were no Cr-Cr bonds in
1
3, then the compound would best be treated as containing three
2+
high-spin Cr ions coupled antiferromagnetically to give an S
1
1
0
3
)
2 ground state as shown in Scheme 8. It might then be
a
Scheme 8
NR ) not reported.
bonds; thus, in the limit that the σ orbitals are totally used for
Ru-C bonding, there is no Ru-Ru σ bond resulting in an
4
2
*4
64
unusual π δ π configuration. For these diruthenium systems,
the Ru-C bond distances are all less than 2.0 Å. Contrarily,
for 9 or 13, the σ bond probably persists because of the long
Cr-C bond distances of ∼2.20 Å.
possible for a higher spin state to be populated at higher
temperatures. However, the magnetic susceptibility data for 2,
9
, and 13 show no indication of this.
Magnetic Susceptibilities. Compound 1 was previously
reported to follow the Curie law with S ) 2.23 The magnetic
This situation may be contrasted with the magnetic suscep-
tibility data for the tricopper compounds Cu3(dpa)4Cl2 and
susceptibility data have been measured for 2, 7, 9, 10, and 13.
They are similar and show Curie behavior with four unpaired
electrons. The plots of øT versus T for these compounds show
1
Cu3(dpa)4(BF4)2 which have S ) /2 ground states arising as in
5
1
Scheme 9.
-
1
a horizontal line at øT ≈ 3 emu K mol , which drops at low
temperatures. An example is shown in Figure 7, and the others
are provided as Supporting Information. This sharp drop in øT
at low temperatures is presumably due mainly to zero-field
splitting of the quintet ground state, but it may also be influenced
by intermolecular antiferromagnetic interactions.
Scheme 9
It is useful to fit 1/ø versus T to the Curie-Weiss equation
for S ) 2. Table 5 summarizes the refined g values and Weiss
constants (θ) for these compounds. Also given in Table 5 are
In both of these compounds, which have Cu‚‚‚Cu separations
of 2.47 and 2.40 Å, respectively, population of the S ) /2 state
3
is apparent above ∼150 K, and the J values from the Hamil-
7+
the refined values for 3 (Cr3 ) based on a fit assuming S )
tonian H ) -J(SA‚SB + SA′‚SB) - J′(SA‚SA′) for these
3
-1
/
2. All of the g values are close to the free-electron value, and
compounds are ∼ -400 cm (J′ values could not be deter-
5
1
the θ values are in the range from -0.13 to -3.2 K. Some of
these compounds (2, 9, and 13) contain symmetrical molecules,
while the rest are unsymmetrical, and one might have expected
that the magnetic behavior would be different for these groups,
since the compounds must have dissimilar electronic structures.
However, they are not, and the two types of species have four
unpaired electrons. The unsymmetical Cr36 chains have smaller
Weiss constants (-0.13 to -0.26 K) than those having
symmetrical Cr36 cores, where θ is at least 3 times as large,
mined).
If symmetrical Cr36 complexes exist with only antiferro-
magnetic coupling between the Cr atoms as in Scheme 8, then
the J values should be similar to those in Cu3(dpa)4Cl2 since
the pathway for coupling should be the same (superexchange
+
6
+
via spin polarization through the dpa ligand). In fact, J for Cr3
+
6+
species would probably be smaller than that for Cu3 species
since, in the latter, the unpaired electrons reside in the dx -y
orbitals which will have very strong interactions with the ligand
2
2
+
7092 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
A R T I C L E S
Table 6. Electrochemical Data for Trichromium Compounds
ligand. There is also a third wave which is very close to the
solvent window which may be assigned to an uncommon
compound
(E1/2)
1
(E1/2)
2
8
+/9+
Cr3
oxidation.
1
2
4
5
7
9
0.067
-0.023
0.853
0.761
not observed
0.57 (irreversible)
0.84
0.9 (irreversible)
0.879
0.795
Replacing the chloride ligands in 1 can drastically change
0.603 (quasi)
0.3 (irreversible)
irreversible
-0.098
6+
the electrochemical behavior of the Cr3 unit, especially with
regard to reversibility. While the CV of 10 with axial thiocyanato
ligands is very similar to that of 1, complexes with more weakly
donating ligands show less reversible behavior. For example, 6
can be oxidized twice at 0.3 and at 0.6 V, but both processes
are irreversible. For complexes having more strongly donating
axial ligands, however, a shift in the first oxidation to lower
potential occurs, as shown in their CVs (Figure 8). Thus, in the
series from Cl to CN to CCPh, the first oxidation potentials for
the complexes plummet from 0.067 V for 1 to -0.098 V for 9
1
1
0
3
0.056
-0.465
a
a
Data from ref 28.
2
8
to -0.465 V for 13. This may be attributed to the ability of
3
+
the axial anion to stabilize the isolated Cr species in the
oxidation product. The second oxidation potentials are not
shifted much from their original position, signifying the inability
5
+
of a more basic axial ligand to stabilize a Cr2 unit. Since CN
and CCPh are known to be exceptional σ donors, these results
suggest that the HOMO of the symmetrical compounds has σ
2
4
symmetry, as suggested by DFT calculations.
Spectroscopy. Frequently, paramagnetic species have unin-
terpretable NMR spectra because the unpaired electrons cause
the nuclear spins to have very short relaxation times, thus
broadening the signals. It is not uncommon for an NMR
spectrum to have no detectable signals for compounds with
unpaired spins. Sometimes an interpretation is possible, however,
Figure 8. Cyclic voltammograms of 1, 2, 9, and 13. The E1/2 values are
given next to each reversible wave, and the data for 13 are from ref 28.
6
6
67
as is the case for Ni3(dpa)4Cl2 and Co3(dpa)4Cl2. We thus
1
decided to study the trichromium EMACs by H NMR
orbitals. Nevertheless, there is no indication in the magnetic
susceptibility data for 2, 9, or 13 that any higher spin excited
state is being populated, and this militates against the possibility
of antiferromagnetic exchange as shown in Scheme 8.
Calculations at the DFT level24 on a symmetrical Cr3(dpa)4Cl2
spectroscopy to see if the spectra could provide information
that would allow differentiation between symmetrical or unsym-
metrical molecular structures in solution.
1
The H NMR spectrum of 1 in CD2Cl2 solution is not readily
interpretable. It features three very broad peaks at -15.2, -1.4,
and 13.6 ppm which have roughly equal intensity, and which
must be due to pyridyl protons. Yet if 1 is symmetrical in
solution there should be four resonances, and if it is unsym-
metrical there should be eight. The fact that there are only three
means that part of the spectrum is unobservable and thus the
structure of 1 cannot be assigned as symmetrical or unsym-
structure also have assumed that the π and δ electrons interact
via antiferromagnetic coupling. Thus, the only bonding interac-
tion taken into account was the 3-center-3-electron σ bond as
discussed above. From these results, the high-spin state (S )
-
1
-1
5
) was calculated to lie 30.8 kcal mol (>10 000 cm ,
-1
corresponding to a J value >500 cm ) above the ground state.
It is unclear to us why such a strong interaction is not considered
bonding, and we believe that a correct description of this system
must at least include π bonding interactions. Perhaps it would
1
metrical. The H NMR spectra of 4, 9, and 10 are similar to
that of 1.
be better to describe the Cr36+ chain (a 12-electron system) as
Interestingly, 2, which is symmetrical in the solid state, shows
1
a H NMR spectrum in which all of the peaks can be assigned.
isolobal to the C3 chain in allene (also a 12-electron system)
with two localized but orthogonal double bonds. We believe
that further theoretical work on this system is needed.
There are only three pyridyl hydrogen atoms in 2, and
resonances corresponding to these appear at 12.2, 24.9, and 27.3
ppm. The ethyl group appears as a triplet at 0.9 ppm for the
CH3 group and a multiplet at 7.2 ppm for the diastereotopic
protons of the CH2 group. This assignment assumes idealized
D4 symmetry for 2 in solution, consistent with what is observed
in the solid state.
7+
Electrochemistry. Oxidation to Cr3 species is an important
6
+
42
process for all Cr3 complexes. Cyclic voltammetry of the
new Cr36 complexes was performed to find out how the identity
of the bridging or axial ligand changes the ability of the Cr3
+
6
+
core to be oxidized. The electrochemical results are shown in
Compounds having a Cr3(dpa)42 core are deeply colored and
dichroic. An intense green hue by reflected light becomes deep
red by transmission through a concentrated solution. The
remarkably complicated electronic spectra in the visible region,
as shown for 1, 2, 4, 9, and 10 in Figure 9, display five distinct
+
Table 6. Compound 1, whose CV is shown in Figure 8, can be
reversibly oxidized twice with E1/2( ) 0.067 V and E1/2
1)
(2)
)
0
.853 V. The first wave is assigned to the oxidation of the
42
isolated Cr(II) species to Cr(III), while the second wave is
4+
tentatively assigned as oxidation of the Cr2 quadruple bond.
As seen for Co3(depa)4Cl241 and Ni3(depa)4Cl2, both waves
for 2 appear at lower potentials than for 1 by ∼0.09 V, a result
attributed to the increased basicity of the ethyl-substituted depa
27
(
66) Cl e´ rac, R.; Cotton, F. A.; Dunbar, K. R.; Murillo, C. A.; Pascual, I.; Wang,
X. Inorg. Chem. 1999, 38, 2655.
(67) Cotton, F. A.; Murillo, C. A.; Wang, X. Inorg. Chem. 1999, 38, 6294.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7093

A R T I C L E S
Berry et al.
spectrum of 1 taken in the solid state is very similar to the
spectrum recorded in solution. In general, quite similar spectra
are obtained in solution for all compounds whether they are
symmetrical or unsymmetrical in the solid state. Thus, the
electronic spectra do not help to distinguish between sym-
metrical and unsymmetrical species. It should also be noted that
1
, although it contains four unpaired electrons, does not show
an EPR spectrum at X-band frequencies at 10 K.
4. Further Comments on the Cr-Cr Bonding in 1-13
Influence of the Axial Ligands. Compounds 1, 4-11, and
1
3 have the general formula Cr3(dpa)4X2 and differ only in the
6+
identity of X. Our previous work indicated that Cr3 complexes
with X ) Cl or CtCPh are symmetrical compounds with
chemically equivalent Cr-Cr bond distances, whereas com-
pounds of the type Cr3(dpa)4XY with X * Y are unsymmetrical
4
+
Figure 9. Electronic spectra of 1, 2, 4, 9, and 10.
with a quadruply bonded Cr2 unit and a long interatomic
2
+
distance to an isolated high-spin Cr ion. This expansion of
our earlier work shows that some of the previous structural
results on different crystalline forms of 1 must be reinterpreted
so that 1 is unsymmetrical in the solid state in every case. DFT
Table 7. Electronic Spectra of Compounds 1, 2, 4, 9, 10, and 13
λ
1
, nm
λ
2
, nm
λ
3
, nm
λ
4
(sh.), nm
-
1
(ꢀ, M cm-1
-1
(ꢀ, M cm-1
-1
(ꢀ, M cm-1
)
-1
compound
(ꢀ, M cm-1
673
1000)
)
)
)
1
(in CH2Cl2)
611
(2000)
612
515
(2000)
520
437
(6000)
24
calculations indicate a symmetrical ground state for 1, which
(
may be a proper description in solution or in the gas phase,
even though it is not in the solid state. Even with the concept
that 1 lies in a potential energy well which is very shallow with
1
2
(solid state)
(in CH2Cl2)
673
664
608
511
436
(
900)
679
2000)
683
2000)
686
1000)
691
1000)
(1000)
561
(1000)
616
(2000)
614
(1000)
625
(2000)
515
(2000)
523
(3000)
512
(2000)
539
(4000)
435
(4000)
455
(7000)
450
(4000)
4
9
1
1
(in CH3CN)
(in CH2Cl2)
24
respect to movement of the central chromium atom, it is
(
difficult to rationalize the trends in the bond distances presented
here, but some important observations can be made.
(
6
+
0 (in CH2Cl2)
3 (in benzene)
All of the unsymmetrical Cr3 complexes can be described
as consisting of a Cr-Cr quadruple bond (dCr-Cr < 2.3 Å) and
an isolated high-spin Cr unit. The bond distances of these Cr-
Cr quadruple bonds fall into three categories:
(
(
(1000)
(2000)
2
.2 Å < dCr-Cr < 2.3 Å (long)
.0 Å < dCr-Cr < 2.2 Å (medium)
dCr-Cr < 2.0 Å (short)
2
From the data in Table 2, it can be seen that good anionic
donor ligands (Cl, NCS, NCO) give rise to long Cr-Cr
quadruple bonds. Medium length Cr-Cr bonds arise from
weaker anionic donor ligands (e.g., iodide) or the neutral
acetonitrile ligands. Very weak σ donors such as nitrate and
tetrafluoroborate give rise to very unsymmetrical complexes with
short Cr-Cr quadruple bonds and an isolated Cr atom which
forms a stronger bond to the axial ligand. The axial ligands
cyanide and phenylacetylide stabilize symmetrical trichromium
chains with equivalent Cr-Cr bond distances >2.35 Å. The
common feature of these ligands is that they are exceptionally
strong σ donors.
To rationalize the Cr-Cr bonding in these different classes
of Cr3(dpa)4X2 compounds in terms of molecular orbital theory,
three major influences must be taken into account. First of all,
three-center bonding must be discussed in relation to the
symmetrical compounds. As shown in Scheme 4, the 3-center-
3-electron σ bond is the most important concept here. Each of
Figure 10. Electronic spectrum in the visible region for 13.
features: two bands of similar intensity (ꢀ ≈ 1000 M- cm )
at ∼670 and 610 nm, a much less intense feature at ∼560 nm,
a more intense band at ∼520 nm, and a very intense band (or
1
-1
-
1
-1
shoulder) at ∼430 nm (ꢀ ≈ 4000 M cm ). The positions of
these bands for 1, 2, 4, 9, 10, and 13 are listed in Table 7. While
these bands cannot be specifically assigned, their intensities and
positions indicate that they are probably d-d transitions. The
symmetrical 13, which is not green but brown, has only three
peaks in the range from 500 to 900 nm at 691, 625, and 539
nm (see Figure 10), which are significantly red shifted from
their corresponding positions in 1, 4, and 10. The visible
the three combinations of σ orbitals (σ, σ , and σ*) can be
nb
displaced in energy by interactions with the axial ligands.
2
4 2
In a Cr-Cr quadruple bond with the configuration σ π δ ,
the strongly bonding σ orbital is destabilized by axial donor
ligands. The more strongly donating the axial ligand is, the
7094 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004

Molecular and Electronic Structures by Design
A R T I C L E S
Table 8. Metal-Metal Separations in Cr3(L)4Cl2 Complexesa
compound
Cr−Cr, Å
Cr‚‚‚Cr, Å
∆dCr-Cr, Å
Cr1‚‚‚Cr3, Å
ref
Cr3(dpa)4Cl2‚benzene
2.227[9]
2.483[9]
2.481[9]
0.256[9]
0.245[9]
0.24[1]
0.223(4)
0.220[4]
0
0.244(1)
0.0222(8)
0.0030(7)
0
4.710[9]
4.719[9]
4.717(2)
4.731(2)
4.717(4)
4.7560(5)
4.782(1)
4.8982(8)
4.9548(7)
4.8478(9)
4.960(1)
4.9486(8)
this work
2
.236[9]
Cr3(dpa)4Cl2‚toluene
Cr3(dpa)4Cl2‚CH2Cl2
Cr3(dpa)4Cl2‚Et2O
Cr3(depa)4Cl2‚0.5hexane
Cr3(PhPyB)4Cl2‚solvents
Cr3(PhPyF)4Cl2‚CH2Cl2
Cr3(AniPyF)4Cl2
2.24[1]
2.48[1]
this work
this work
this work
this work
58a
2.254(4)
2.249[4]
2.3780(5)
2.269(1)
2.4380(8)
2.4759(7)
2.4298(8)
2.460(1)
2.4743(8)
2.477(4)
2.469[4]
symmetrical
2.513(1)
2.4602(8)
2.4789(7)
symmetrical
2.500(1)
58b
58b
Cr3(TolPyF)4Cl2‚2H2O
58b
F
Cr3(Ph PyF)4Cl2
0.040(1)
0
58b
cis-Cr3(PhPcF)4Cl2‚THF‚
symmetrical
58b
0
.5hexane
(
3:1)-Cr3(PhPcF)4Cl2‚
2.216(1)
2.646(1)
0.430(1)
4.862(1)
58b
THF‚0.61Et2O
a
Brackets represent average values.
longer the Cr-Cr bond distance becomes. In the unsymmetrical
trichromium complexes, the Cr-Cr bond has an axial ligand
on one side only as shown in Scheme 10.
strongly donating ligands will completely break the Cr-Cr bond,
but no effort to make such compounds has been reported so
the point is moot.
It should also be mentioned here that by employing other
Scheme 10
6+
bridging ligands several compounds with Cr3 chains have been
5
9
synthesized which have no axial ligands at all. As in the case
of complexes with very weak donor ligands, these are all
unsymmetrical with very short Cr-Cr quadruple bonds.
Influence of the Equatorial Ligands. Besides the ligand dpa,
previous work in this laboratory has employed other polydentate
bridging equatorial ligands to stabilize trichromium compounds
of the type Cr3(ligand)4Cl2. Some have been unsymmetrical
These complexes contain, however, another unit which must
be taken into account: an isolated square pyramidal Cr unit,
which has the orbital diagram shown in Scheme 11.
II
58
amidinates such as those shown in Scheme 12.
Scheme 12
Scheme 11
Compound 2, Cr3(depa)4Cl2, is the first Cr36 species in which
the bridging ligand is a modification of the dipyridylamido
ligand. Since all of these compounds (summarized in Table 8)
have axial chloride ligands, it is useful to compare them to
determine the role of the bridging ligand in the Cr-Cr bonding.
As shown in Table 8, the compounds with unsymmetrical
formamidinate ligands tend to crystallize with symmetrical
structures. The small differences in the two Cr-Cr distances,
∆dCr-Cr, despite being statistically meaningful, are chemically
insignificant. An exception is (3:1)-Cr3(PhPcF)4Cl2, in which
the unsymmetrical Cr3 chain is thought to derive from the
unsymmetrical 3:1 arrangement of the formamidinate ligands.5
By changing from formamidinate ligands to a benzamidinate
+
In all cases presented here, the isolated unit is high spin,
giving rise to an S ) 2 ground state for the complexes, and
strongly donating axial ligand orbitals of a symmetry (in the
group C4) will destabilize the dz orbital.
Thus, the influence of more strongly donating axial ligands
2
is three-fold: strong axial σ donors will destabilize both the
2
+
Cr-Cr quadruple bond and the isolated high-spin Cr ion by
2
raising the Cr-Cr σ bond (shown in Scheme 10) and the dz
orbital (shown in Scheme 11) in energy. On the other hand, for
a delocalized Cr3 chain, these axial ligands will destabilize the
σnb orbital and to some extent the σ and σ* orbitals, and this
lengthens the Cr-Cr bonds but does not destroy the 3-center
bonding in the complex. Thus, very weakly donating axial
8b
6
+
ligand in Cr3(PhPyBz)4Cl2, the Cr3 chain becomes distinctly
unsymmetrical, with a geometry similar to those unsymmetrical
structures of Cr3(dpa)4Cl2. The ethyl-substituted Cr3(depa)4Cl2
is also symmetrical, whereas Cr3(dpa)4Cl2 is not.
6+
ligands favor an unsymmetrical Cr3 chain with short Cr-Cr
quadruple bonds. Stronger donors may also give rise to
unsymmetrical chains but with a correspondingly longer Cr-
Cr quadruple bond. Very strong donors completely destroy the
localized Cr-Cr quadruple bond, and the molecule responds
to this by employing 3-center bonding which is not destabilized
to the same degree as the localized Cr-Cr σ bond. It should be
One major factor affecting the occurrence of a symmetrical
versus an unsymmetrical structure is the basicity of the bridging
ligands. We have noted before that the ligand depa is more basic
4
1
than dpa, because 4-ethylpyridine is more basic than pyridine
6
8
by 0.8 pK units. Also, it is reasonable to assume that the
phenylpyridylbenzamidinate ligand, PhPyBz, is less basic than
phenylpyridylformamidinate, PhPyF, due to the electron-
4+
mentioned that no simple Cr2 complex is known with axial
cyanide or acetylide ligands.11 Perhaps the reason is that these
J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004 7095

A R T I C L E S
Berry et al.
+
withdrawing influence of the central phenyl group. Thus, more
each crystalline compound consists of a quadruply bonded Cr24
6+
2+
basic bridging ligands appear to favor symmetrical Cr3 chains,
whereas less basic ligands favor unsymmetrical compounds. This
is probably due to the interaction of the π orbitals of these
ligands with the δ orbitals of the trichromium unit, similar to
the interactions between ligand π orbitals and M-M δ orbitals
that lead to the unique properties in complexes of the ligand
unit (Cr-Cr ) 2.24 Å) and an isolated high-spin Cr species
at a longer distance of 2.48 Å. Thus, reported DFT calculations24
on a symmetrical model of 1 may have been performed on a
hypothetical molecule, although a symmetric structure of 1 in
solution or in the gas phase cannot be ruled out.
Further insight into the bonding in these trichromium com-
pounds has arisen through the study of the properties of a series
of Cr3(dpa)4X2 compounds where the X anions have been weak,
intermediate, and strong σ donors: BF4, NO3, NCCH3, Cl, Br,
I, NCS, NCO, CN, and CCPh. Use of most of these ligands
yields unsymmetrical complexes like 1, but use of the excep-
tionally strong donors CN and CCPh result in Cr3(dpa)4X2
complexes (9 and 13) with symmetrical Cr3 chains. Thus, the
Cr-Cr bond distances can be tuned by choice of the axial ligand.
The range of Cr-Cr distances which can be thus obtained is
very large (1.93-2.43 Å), although all of the compounds show
similar properties. Oddly, the Cr-X bond lengths in this series
are uniformly longer than the M-X distances in corresponding
6
9
1
,3,4,6,7,8-hexahydro-2H-pyrimido[1,2-a]pyrimidinate.
Thus, in an unsymmetrical molecule, a more basic ligand will
4
+
destabilize the localized δ bond of the quadruply bonded Cr2
unit through antibonding interactions of the filled π orbitals of
the ligand of b symmetry in the C4 point group with the δ bond.
However, in a symmetrical molecule (D4 symmetry), the δ
orbitals are essentially nonbonding, so no destabilization of the
Cr-Cr bonds occurs. Thus, more basic bridging ligands will
destabilize the Cr24+ unit of an unsymmetrical compound and
cause a symmetrical structure to be favored.
7+
Oxidation to Cr3 Species. All of the previously reported
Cr37 compounds, whether they contain either the [Cr3(dpa)4-
+
+
M (dpa) X compounds (M ) Ni, Co) by as much as 0.3 Å.
Cl2] cation or cations with radically different axial ligands on
3
4
2
We believe these geometrical features to be clues to the correct
description of the electronic structure of these compounds and,
more importantly, of all Cr-Cr bonded compounds, which have
thus far only frustrated theoreticians. The nature of the bridging
ligands also has an important impact on the electronic structure
of the trichromium chains. The more basic depa ligand was used
to synthesize Cr (depa) Cl , 2, which has equal Cr-Cr distances.
Thus, there are two main factors which allow the design of a
symmetrical Cr3 chain. One is to embrace the Cr atoms with
very basic bridging ligands. Another is to employ strong σ-donor
axial ligands.
Electrochemical studies of the Cr36 complexes have shown
that as ∆dCr-Cr (i.e., the degree of asymmetry) increases, the
oxidations to Cr3 become less reversible. The symmetrical 2,
9, and 13 show reversible Cr3 /Cr3 redox couples which
become more accessible as the σ donor strength of the axial
ligand increases. This suggests that the HOMO of the sym-
either end like compound 12, are distinctly unsymmetrical. Yet
since Cr3(dpa)4XY compounds are generally very unsymmetrical
and we have shown here that neutral Cr3(dpa)4Cl2 is also
unsymmetrical, the question remains: what happens when a
17
symmetrical Cr36 complex is oxidized to Cr3 ? Will the
+
7+
molecule remain symmetrical, or will it become unsymmetrical,
+
like the [Cr3(dpa)4Cl2] cations? If the latter is true, then these
3
4
2
complexes may find useful applications as molecular switches,
since a symmetrical molecule, presumably, would be a better
conductor than an unsymmetrical one.
6
+
With this question in mind, 2 was oxidized to 3 with
ferrocenium triiodide. The structure of 3 is quite unsymmetrical
with ∆dCr-Cr ) 0.295(8) Å and the short Cr-Cr bond distance
of 2.146(8) Å. This is slightly longer than the Cr-Cr distance
+
7
+
6
+
7+
+
in the [Cr3(dpa)4Cl2] cation (2.07[1] Å), but nevertheless shows
that the electronic structure of Cr3(depa)4Cl2 changes upon
oxidation from a delocalized situation to one having a localized
6
+
4
+
III
metrical Cr3 chain is σ in character, which is in agreement
Cr2 quadruple bond and an isolated high-spin Cr ion. We
6
+
with the results of DFT calculations. Since oxidation of a
believe that this will be the case for all symmetrical Cr3
complexes.
6
+
7+
complex with a symmetrical Cr3 core to one with a Cr3
core had not previously been reported, 2 was oxidized to 3
5. Conclusions
+
-
7+
(
2 I3 ), wherein an unsymmetrical Cr3 chain was found (along
-
6+
with an unsymmetrical I3 chain). Thus, symmetrical Cr3
This paper presents a wealth of data about compounds
complexes could be used as molecular switches which can be
turned on or off via an applied potential.
containing linear trichromium chains. The crystal structures of
solvates of Cr3(dpa)4Cl2, 1, which were previously reported to
be symmetrical are here shown to be correctly described as
unsymmetrical. The change in description results from a metic-
ulous crystallographic analysis of disorder of unsymmetrical
molecules of C4 symmetry which pack in such a way that with
single Cr atoms placed (incorrectly) in their averaged positions
the molecule appears to have D4 symmetry. The geometries of
all crystalline forms of 1 (except for the twinned 1‚THF) are
now consistent with each other, and the Cr-Cr distances are
essentially the same for each polymorph. Furthermore, previous
reports that 1 is symmetric in certain solid-state environments
are incorrect. Present results show that the molecule of 1 in
Acknowledgment. We thank the NSF for support through a
Nanoscale Science and Engineering/NIRT Grant (DMR 0103455)
and the Telecommunications and Information Task Force of
Texas A&M University. We thank Prof. N. S. Dalal for allowing
B.K.R. to participate in this work while training in synthesis at
the Laboratory for Molecular Structure and Bonding. J.F.B.
thanks the NSF for support in the form of a predoctoral
fellowship. We also thank Dr. S. Silber and Dr. V. I. Bakhmutov
for helpful discussions on NMR.
Supporting Information Available: Crystallographic data in
CIF format and magnetic susceptibility data in PDF format for
2
, 3, 9, and 10. This material is available free of charge via the
(
68) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum Press:
New York, 1982.
Internet at http://pubs.acs.org.
(69) Cotton, F. A.; Gruhn, N. E.; Gu, J.; Huang, P.; Lichtenberger, D. L.; Murillo,
C. A.; Van Dorn, L. O.; Wilkinson, C. C. Science 2002, 298, 1971.
JA049055H
7096 J. AM. CHEM. SOC.
9
VOL. 126, NO. 22, 2004