Bond distances are not always what they appear to be: Discovery and un-discovery of the longest Cr(v)≡N triple bond

Article abstract of DOI:10.1039/b717618j

The coordination chemistry of CrCl₃ with three pentadentate ligands having the [1,4,7]-triazacyclononane-1,4-diacetato motif have been investigated. The new resulting six-coordinate Cr(iii)-chloro species react cleanly with sodium azide to form the corresponding azido species, which undergo photolysis under irradiation at 419 nm in acetonitrile-water solution to form Cr(v)-nitrido species that are partially hydrolyzed to their corresponding Cr(iii)-hydroxo counterparts. Five of these Cr complexes have been characterized by X-ray crystallography. The hydroxo and nitrido species co-crystallize complicating crystal structural refinement. What at first appeared to be the longest Cr≡N triple bond distance yet observed (1.66 A) was found after re-refinement to be an artifact of positional disorder of the Cr atoms of 73% nitrido and 27% hydroxo species. The re-refined Cr≡N bond distance is estimated as 1.58 A, which agrees well with the observed Cr≡N stretching frequency of 971 cm^-1 found by IR spectroscopy. UV-vis data for the nitrido complexes suggest that they are all three partially hydrolyzed. These results emphasize the care that must be taken in the characterization of compounds that may co-crystallize with structurally similar analogs. The Royal Society of Chemistry.

Full text of DOI:10.1039/b717618j

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Bond distances are not always what they appear to be: discovery and
un-discovery of the longest Cr(V)≡N triple bond†‡
a
Yu-Fei Song,*§ John F. Berry,*b Thomas Weyhermu¨llera and Eckhard Billa
Received 14th November 2007, Accepted 21st December 2007
First published as an Advance Article on the web 16th January 2008
DOI: 10.1039/b717618j
The coordination chemistry of CrCl3 with three pentadentate ligands having the
[1,4,7]-triazacyclononane-1,4-diacetato motif have been investigated. The new resulting six-coordinate
Cr(III)-chloro species react cleanly with sodium azide to form the corresponding azido species, which
undergo photolysis under irradiation at 419 nm in acetonitrile–water solution to form Cr(V)-nitrido
species that are partially hydrolyzed to their corresponding Cr(III)-hydroxo counterparts. Five of these
Cr complexes have been characterized by X-ray crystallography. The hydroxo and nitrido species
co-crystallize complicating crystal structural refinement. What at first appeared to be the longest Cr≡N
˚
triple bond distance yet observed (1.66 A) was found after re-refinement to be an artifact of positional
disorder of the Cr atoms of 73% nitrido and 27% hydroxo species. The re-refined Cr≡N bond distance
−1
˚
is estimated as 1.58 A, which agrees well with the observed Cr≡N stretching frequency of 971 cm
found by IR spectroscopy. UV-vis data for the “nitrido” complexes suggest that they are all three
partially hydrolyzed. These results emphasize the care that must be taken in the characterization of
compounds that may co-crystallize with structurally similar analogs.
common features (1) they are five-coordinate, having square-
Introduction
˚
pyramidal geometry; (2) the Cr≡N distance is around 1.56 A
and the CrV ion lies ∼0.5 A above the plane defined by the
High-valent chromium species have long been of general interest
as powerful oxidants for organic synthesis.1 For example, Cr(V)-
oxo species are well known to promote epoxidation of olefins,
and chromic acid (Cr(VI)) is often used to attack activated
hydrocarbon C–H bonds.2 Corresponding high-valent chromium-
nitride chemistry is less well developed. The field began in
1981 with the postulation of a Cr(V)-nitrido complex resulting
from photolysis of [CrIII(salen)(N3)]·2H2O.3 In 1983, the crystal
structure of [CrV(N)(tpp)] (tpp = tetra-tolylporphinate(2-)) was re-
ported by Groves et al.,4 providing the first definitive proof for the
formation of a Cr≡N moiety. Subsequently, other ligands such as
1,2-bis(2-pyridylcarboxy)-amido)benzene5 and phthalocyaninate6
were employed. Recently, a very efficient synthetic “tool” for
the preparation of Cr(V)-nitrido complexes was developed by
Bendix using atom and/or group transfer reactions,7,8 which have
allowed the synthesis of Cr≡N species with mono-, or bidentate
ligands, instead of the polydentate ligands necessary in earlier
syntheses. Notably, the above compounds have the following
˚
four equatorial atoms of the ligand; (3) the m(Cr≡N) stretching
frequency is observed at ∼1015 cm−1.
In 1996, the first distorted octahedral Cr(V)-nitrido complexes
were reported,9 of the type [(tacn)Cr(N)L]n+ where tacn is 1,4,7-
triazacyclononane and L is a bidentate ligand such as oxalate,
acetonylacetonate, 2-picolinate or o-phenanthroline. Later, other
six-coordinate Cr(V) complexes derived from the ligand 1,4,8,11-
tetraazacyclotetradecane (cyclam) were introduced.10 These com-
plexes differ from the five-coordinate Cr(V)-nitrido complexes in
that the Cr≡N stretching frequency is observed at lower frequency
∼967 cm−1. The mean Cr≡N bond distance for the six-coordinate
˚
species, 1.57 A, is only slightly longer than in the case of the five-
coordinate species.
We report here some surprising results on Cr(V)-nitrido species
supported by pentadentate derivatives of tacn. Such ligands,
which are not as well represented as mono or tri-substituted tacn
systems, are currently of interest because these macrocycles can
be utilized to construct mononuclear octahedral complexes that
possess a site for binding and activation of small molecules.11–14
We have recently reported three new ligands containing the
1,4,7-triazacyclononane-1,4-diacetate motif (Chart 1) and their
coordination chemistry with iron,15 and now report new chromium
complexes with these ligands. In this paper, nine octahedral
chromium compounds with the ligands shown in Chart 1 have
been prepared and characterized. Several of these compounds have
been structurally characterized, and we also report a non-trivial
crystal structure of a new Cr(V)-nitrido complex, that crystallizes
readily with the analogous Cr(III) hydroxo complex, causing an
aMax-Planck-Institut fu¨r Bioanorganische Chemie, Stiftstrasse 34-36, D-
45470 Mu¨lheim an der Ruhr, Germany. E-mail: songyufei@hotmail.com
bDepartment of Chemistry, University of Wisconsin - Madison, 1101
University Ave., Madison, WI, 53706, USA. E-mail: berry@chem.wisc.edu
† CCDC reference numbers 667804–667808. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/b717618j
‡ Electronic supplementary information (ESI) available: Details of the
crystal data collection, atom coordinates, anisotropic thermal parameters
and calculated positional parameters for the hydrogen atoms and bond
distances and angles. See DOI: 10.1039/b717618j
§ Present address: Department of Chemistry, University of Glasgow,
Glasgow, UK G12 8QQ.
1864 | Dalton Trans., 2008, 1864–1871
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Crystal structures
Crystal structures of one chloro complex, 2, all three azido com-
plexes, 4–6, and one Cr(V)-nitrido complex, 9, were determined
by single crystal X-ray crystallography. A summary of structural
parameters is given in Table 1 and selected bond lengths are shown
in Table 2, except those of compound 9, which will be presented
later.
Fig. 1 shows views of 2 and 6; structures of 4 and 5 are shown
in the electronic supplementary information (ESI).‡ The ligand
arrangement in each crystal structure is similar. The coordination
spheres of the Cr centers are facially capped by the three N atoms
of the tacn backbone with two of the three further coordination
sites filled by the O atoms of the appended acetate groups and the
sixth site occupied by Cl in the case of 2 or azide for 4–6. There
is a slight trigonal distortion to all of the complexes as◦evidenced
Chart 1
unusually long Cr≡N bond distance to appear in the incompletely
refined crystal structure, similar to effects first analyzed by Parkin
for other metal–ligand multiply bonded systems.16
Discovery of the longest Cr≡N triple bond distance
◦
by the deviation of the X–Cr–Y bond angles from 90 by ∼5 ,
Preparation of complexes
where X and Y are cis donor atoms to the Cr center. This trigonal
distortion coincides with the simplified trigonal geometry of the
tacn backbone, and was also observed in the corresponding iron
complexes.15
The Cr–N bond distances in 2 and 4–6 follow an interesting
trend. The Cr–N bond to the alkyl or aryl N atom has a mean
The Cr(III) complexes of L1, L2, and L3 are conveniently
synthesized by reaction of stoichiometric amounts of the ligand
and Cr(THF)3Cl3 in refluxing methanol. Reaction of the so-
formed Cr(III)-chloro species (compounds 1, 2, and 3 having
ligands L1, L2, and L3, respectively) with excess NaN3 furnishes
the corresponding Cr(III)-azido species (compounds 4, 5, and
6). Solutions of these deep purple complexes in methanol–water
slowly undergo photochemical cleavage (with 419 nm light) of the
azido moiety yielding light pink Cr(V)-nitrido complexes (7, 8, and
9). The photolysis process was monitored optically by changes in
the color of the species from dark purple to light pink. A solid
sample of 9 was prepared; it was established gravimetrically that
˚
distance of 2.1223[8] A that is significantly longer than the other
˚
˚
two Cr–N distances, 2.0614[8] A and 2.0706[8] A for the Cr–N
distances cis and trans to the monodentate Cl or N3 ligand, which
are statistically different from each other, though the difference
is unlikely to be relevant in a chemically meaningful sense. The
˚
Table 2 Selected bond distances (A) for 2, 4, 5, and 6
1.0
0.2 equiv of N2/Cr(III) center evolved from the starting
2
4
5
Cr1–N7 2.0610(9)
Cr1–O16 1.9629(8)
Cr1–N4 2.0789(9)
Cr1–O20 1.9840(8)
Cr1–N1 2.1238(9)
Cr1–Cl1 2.3082(3)
azido-Cr(III) complex. The reactions are summarized in Scheme 1.
Cr1–O22 1.950(1)
Cr1–N7 2.052(2)
Cr1–O26 1.953(1)
Cr1–N4 2.073(2)
Cr1–N31 2.006(2)
Cr1–N1 2.112(2)
Cr1–N31 2.021(2)
Cr1–N1 2.125(2)
Cr2–O60 1.963(2)
Cr2–N47 2.058(2)
Cr1–N7 2.067(2)
Cr1–O16 1.964(2)
Cr2–O56 1.964(2)
Cr2–N44 2.073(2)
Cr1–N4 2.073(2)
Cr1–O20 1.951(2)
Cr2–N71 1.984(2)
Cr2–N41 2.132(2)
6
Cr1–O22 1.970(2)
Cr1–N7 2.061(2)
Cr1–O26 1.952(2)
Cr1–N4 2.063(2)
Cr1–N31 2.018(2)
Cr1–N1 2.119(2)
Scheme 1
Table 1 Crystal data
2
4
5
6
9
Chemical formula
FW
C13H27ClCrN3O6
408.83
Monoclinic
P21/c
6.8717(3)
31.2767(12)
8.6425(4)
108.478(4)
1761.72(13)
4
C17H23CrN6O4
427.41
Orthorhombic
P212121
7.4667(6)
12.8766(9)
18.77939(9)
90
1805.5(2)
4
C13H24CrN6O4.5
388.38
Monoclinic
P21/n
7.2406(2)
27.785(2)
16.5271(6)
96.459(5)
3303.8(3)
8
C18H25CrN6O5
457.44
C18H31.3CrN3.7O8.3
484.29
Monoclinic
C2/c
33.841(1)
6.7077(2)
18.7774(4)
97.07(3)
4230.0(2)
8
Crystal system
Space group
Orthorhombic
Pbca
13.2255(6)
9.4894(4)
30.5577(12)
90
3835.1(3)
8
1.585
˚
a/A
˚
b/A
˚
c/A
b/◦
3
˚
V/A
Z
qcalcd/g cm−1
1.541
1.572
1.562
1.521
Refl. collected/2hmax
No. of params/restr.
R, wR2 (I > 2r(I))
R indices (all data)
54167/31.06
231/6
0.0257, 0.0585
0.0342, 0.0619
38954/33
253/0
0.0475, 0.0796
0.0635, 0.0847
42349/30
452/1
0.0521, 0.0962
0.0843, 0.1077
38769/30
272/0
0.0577, 0.0986
0.0848, 0.1071
38270/31.07
290/18
0.0730, 0.1215
0.1044, 0.1314
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1864–1871 | 1865
©

View Article Online
Table 3 Observed absorption maxima from UV-vis measurementsa
Complexes
k/nm (e/L mol−1 cm−1
)
4
5
6
7
8
9
540 (190), 400 (152), 317 (sh), 268 (4.1 × 103)
548 (172), 403 (118), 310 (sh), 273 (3.0 × 103)
537 (200), 405 (120), 320 (sh), 268 (4.5 × 103)
565 (55), 420 (41), 330 (sh), 270 (sh)
567 (46), 427 (42), 328 (sh), 278 (sh)
555 (49), 438 (41), 345 (sh), 275 (sh)
a sh: shoulder.
unpaired electron. This data is in notable agreement with the
formulation of 9 as a nitrido complex of the d1 Cr(V) ion.
Table 3 summarizes the UV-vis spectral data, measured in
MeCN–H2O (1 : 1) at ambient temperature. In the case of the
azido-Cr(III) complexes in roughly octahedral symmetry, two d–d
transitions of low intensity in the visible region at ∼540 and 400 nm
4
4
4
4
are assigned as A2g→ T2g(F) and A2g→ T1g(F), from which the
ligand field splitting energy 10Dq is calculated to be 18 500 cm−1,
18 200 cm−1, and 18 600 cm−1 for 4, 5, and 6. Accordingly,
the Racah parameter B may be estimated from the strong-field
approximation without taking into account of the interaction
between 4T1g(F) and the unobserved 4T1g(P). This gives the values
540 cm−1, 550 cm−1, and 510 cm−1 for 4, 5, and 6, respectively.
The value of 10Dq is therefore largest for the complex bearing
the most electron-donating R group, p-CH3OC6H4, though the
effect of other R groups does not significantly affect the electronic
structure.
Fig. 1 Thermal ellipsoid plots of (A) 2 and (B) 6 drawn at the 50%
probability level. Hydrogen atoms have been removed for clarity.
˚
reason for the ∼0.05 A elongation of the alkyl or aryl Cr–N
bond distance is not readily apparent, but may be explained by
a necessary shortening of the other two Cr–N bond distances to
allow chelation of the N-acetate groups. Other bond distances to
chromium fall within expected ranges.
Our initial refinement of the structure of the nitrido complex 9
showed the expected connectivity of the atoms, having the ligand
L3 in the same coordination mode as seen in all of the other
structures. The most prominent differences between this structure
and the other reported structures are the presence of a short Cr–
Compounds 7, 8, and 9 display d–d transitions at ∼560 nm,
∼430 nm and ∼330 nm of low intensity, and assignment of these
bands will be discussed later (Fig. 2).
˚
N bond distance of ∼1.66 A, tentatively assigned as a Cr≡N
nitrido unit. In agreement with this formulation is the fact that
the Cr atom lies significantly above the mean equatorial N2O2
˚
plane by 0.28 A and that the Cr–N distance trans to the putative
˚
nitrido ligand, 2.24 A, is significantly long as compared to the
comparable distances in 2 and 4–6. As discussed above, the average
Cr≡N bond lengths in five-coordinate Cr(V)-nitrido species are
˚
˚
∼1.56 A, and those of six-coordinate species are ∼1.57 A. The
˚
apparent Cr≡N bond distance in 9, 1.66 A, is therefore quite
puzzling. If this distance were correct, it would be the longest
Cr≡N bond distance known, even longer than that found in
the [Cr(N)(CN)5]3− ion, 1.594(9) A, having a cyano group trans
˚
to the nitrido ligand.17 Further evidence confirming the Cr≡N
connectivity and confirming its length is therefore necessary.
Further characterization
Fig. 2 UV-Vis spectra of 6 during photolysis to produce 9.
Magnetic susceptibilities and electronic spectra of complexes
Magnetic susceptibilities of solid samples were measured by
using a SQUID magnetometer. For the mononuclear azido-Cr(III)
complexes, temperature-independent magnetic moments of 3.84
lB, 3.80 lB and 3.78 lB have been observed at 10–300 K for
4, 5, and 6, respectively, which are in good agreement with
the d3 electron configuration of the Cr(III) ion. A temperature-
independent magnetic moment of 1.85 lB at 30–290 K has been
found for compound 9 in agreement with a species having one
EPR spectroscopy
X-Band EPR spectra of the azido-Cr(III) complexes were mea-
sured at 10 K in CH3CN–H2O (1 : 1) solution. Rhombic signals at
gy = 4.8, gx = 2.8 and gz = 1.9 were observed with broad linewidths,
typical for Cr(III) complexes (d3 electronic configuration, 4F
ground state in Oh symmetry) of rather low symmetry.
1866 | Dalton Trans., 2008, 1864–1871
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
in Table 4 to those of other six-coordinate and five-coordinate
Cr(V)-nitrido complexes. While the Cr≡N stretching frequency in
9 of 971 cm−1 is lower in energy than many of the other known
Cr(V)-nitrido species, it is not as low as would be expected based
˚
on an apparent Cr≡N bond distance of 1.66 A. For example,
the compound [(Me3tacn)CrN(acac)]ClO4 has a Cr≡N stretching
−1
9
˚
frequency of 967 cm , and a Cr≡N bond distance of 1.58 A.
Un-discovery of the longest Cr≡N triple bond distance
Computations
We were initially puzzled by the facts that the crystal structure
of 9 contained such a long Cr≡N bond distance whereas other
physical and spectroscopic data for 9 seemed to indicate that
there was nothing unusual about the Cr≡N species. Unable to
rationalize such a long Cr≡N distance, we therefore decided to
model the structure of 9 using computational methods at the
density functional level of theory (DFT). For the computational
model, the ligand L3 was truncated to the NH analog as shown in
Scheme 2.
Scheme 2
Fig. 3 X-Band EPR spectra of (A) 8 and (B) 9 in 1 : 1 CH3CN–H2O
solution at 293 K. Microwave frequency: 9.4334 GHz, power: 5.029 lW,
modulation amplitude: 1.2 mT.
Using this truncated model ligand, the corresponding LCrN
complex was subjected to a geometry optimization using the BP86
functional, which has been previously used to compute accurate
geometries of species having metal–nitrogen triple bonds.19 The
results of the geometry optimization of this model system are
shown in Table 5 in comparison to the apparent metal–ligand
bond distances in the structure of 9. Notably, the DFT optimized
geometry disagrees with the crystallographic result in that a
In contrast, the EPR spectra of frozen solutions of 7, 8, and
9 in MeCN–H2O (1 : 1) at 10 K showed a broad unresolved line
at giso = 1.98 without any g or hyperfine splitting, presumably
due to significant intermolecular interactions. When EPR spectra
of fluid solutions were measured at room temperature (Fig. 3),
narrow isotropic signals were detected with the typical quartet of
hyperfine satellite lines arising from the 9.5% abundant 53Cr with
I = 3/2. Isotropic g values of 1.979, 1.977, 1.977 and coupling
constants, Aiso(53Cr), of 2.679, 2.705 and 2.647 mT were obtained
from simulations for complexes 7, 8, and 9, respectively. No
hyperfine splitting from interaction with 14N nuclei (I = 1) was
resolved for any samples, possibly due to motional averaging of
anisotropic couplings to the coordinated ligands.18 Thus the EPR
spectra of 7, 8, and 9 resemble those of previously published Cr(V)-
nitrido species.18
˚
Cr≡N bond distance of 1.55 A is predicted by DFT instead of
˚
the apparently observed 1.66 A bond distance. Another striking
difference in the two structures is that the DFT result predicts a
˚
Cr–N bond distance of 2.44 A for the nitrogen atom trans to the
˚
nitrido ligand, whereas the observed distance is 2.24 A. Because
of these computational results, we began to doubt the accuracy
of the crystal structure of 9 and initiated a reinvestigation of the
crystallographic data.
Reinvestigation of the crystal structure of 9
Upon closer inspection of the crystal structure of 9, we noted that
the thermal ellipsoid for the Cr atom was slightly elongated in
the direction of the Cr≡N bond (Fig. 4). This unusual feature
suggests the possibility that the Cr atom may be disordered
over two crystallographically independent positions that are close
enough to each other that the “average” Cr position appears
somewhere in between, and the actual electron density of the Cr
centers is covered by the elongation of the thermal ellipsoid of the
“average” Cr atom.20,21 On the other hand, the thermal ellipsoid
for the nitrido nitrogen atom is not prolate, nor is the ellipsoid
Infrared spectroscopy
The formation of a Cr(V)≡N group should be detectable by
infrared (IR) spectroscopy, and, indeed, photolyzed samples of 7,
8, and 9 show a new band in their IR spectra at ∼971 cm−1 that may
be assigned to a Cr≡N stretch. Also, the strong absorption bands
at 2060 cm−1 of the azido complexes due to the N3 unit decreased
dramatically during photolysis and disappeared completely in the
end. The assigned Cr≡N stretching frequencies are compared
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1864–1871 | 1867
©

View Article Online
Table 4 Comparison of known Cr(V)-nitrido speciesa
m(Cr≡N)/cm−1
Ref.
˚
˚
Complexes
C.N.
Cr≡N/A
Cr out of plane/A
Cr(N)(tpp)
Cr(N)(bpb)
Cr(N)(salen)
Cr(N)(bis-8-hql)
Cr(N)(bis-dppd)
Cr(N)(bis-pddtc)
[Cr(N)Cl4]2−
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
1.565(6)
1.560(2)
1.544(3)
1.5549(7)
1.5491(14)
1.5490(13)
1.56(2)
1.545(2)
1.541(4)
1.575(9)
1.561(3)
1.570(3)
1.594(9)
1.563(2)
0.42
0.51
0.499(3)
0.5193(3)
0.4652(6)
0.7251(13)
1017
1015
1012
1015
1007
991
1009
1016
4
5
3
8
8
8
7
18
18
9
10
33
17
18
[Cr(N)(NCS)4]2−
[Cr(N)(N3)4]2−
[(Me3tacn)Cr(N)(acac)ClO4]
trans-[Cr(N)cyclam(MeCN)]
trans-[Cr(N)(CN)4(py)]
Cr(N)(CN)5
Cr(N)(NCS)3(Phen)
7
8
9
0.41
0.53
0.30
0.265
0.261(2)
1.594(9)
0.28
1019
IR: not detected; Raman: 967
996
995
995
970
971
968
971
This work
This work
This work
See
Table 5
a 8-hql: 8-hydroxoquinolinate; dppd: 1,3-diphenylpropane-1,3-dionate; pddtc: pyrrolidinedithiocarbamate.
Table 5 Structural parameters for 9
Initial crystal structure Re-refined crystal structure DFT
˚
Cr≡N/A 1.666(2)
1.585(8)
2.118(2)
2.059(2)
2.286(2)
1.984(2)
1.999(2)
1.553
2.148
2.153
2.437
1.941
1.939
˚
Cr–N1/A 2.116(2)
˚
Cr–N2/A 2.061(2)
˚
Cr–N3/A 2.238(2)
˚
Cr–O1/A 1.978(2)
˚
Cr–O2/A 1.978(2)
for the nitrogen atom trans to the nitride. Thus, if the Cr atom is
disordered over two positions, there are necessarily two different
sets of Cr–N bond distances. Such a disorder model was found by
Parkin16 to explain unusual bond lengths in solid solutions, most
commonly due to disorder of the ligand positions. Also relevant
is recent work on metal–metal multiply bonded species in which
the metal atom sites are disordered, initially leading to unusual
metal–metal bond distances.20,21 The most likely explanation for
the unusual crystallographic parameters in 9 is that 9 is probably
co-crystallized with a second compound that does not have a short
Cr≡N bond. The most likely possibility for such a compound
would be a chromium(III) hydroxo complex, LCrOH, that can be
formed by hydrolysis of the Cr≡N species.
Fig. 4 Crystal structure of 9 (A) before and (B) after re-refinement of
the disordered Cr atom. Displacement ellipsoids are drawn at the 50%
probability level and hydrogen atoms as well as the p-C6H4OCH3 group
have been removed for clarity.
Further evidence for a disordered Cr atom in 9 is seen from
a plot of its electron density. When the Cr atom is refined using
anisotropic displacement parameters, its electron density in the
N31–N4–N7–O26–Cr mean plane (Fig. S3‡) shows significant
elliptical distortion. Moreover, isotropic refinement of the Cr atom
leads to the appearance of two “lobes” of electron density on either
side of the Cr atom that are visible in the difference map (Fig. S4‡).
A disorder model assuming co-crystallization of 9 with LCrOH
were constrained to have the same coordinates and displacement
parameters. When free refinement of this structural model failed
(the two Cr atoms coalesced into one again), a constraint was
˚
used fixing the Cr≡N bond distance at 1.55 A (suggested from
the computational result). Addition of this constraint allowed the
refinement to converge, and once convergence was achieved, the
˚
was tested by creating two new Cr positions, about 0.3 A from each
˚
other along the Cr≡N axis, whose occupancies were constrained
to sum to 1, and splitting the nitrido N atom into an N atom and an
O atom whose occupancies were constrained to be equal to the Cr
atom having the short Cr≡N distance and the Cr atom having the
longer Cr–O distance, respectively. The metal atoms were refined
with isotropic displacement parameters and the N and O atoms
1.55 A constraint was removed and the Cr≡N bond distance was
˚
refined and converged to a value of 1.585(8) A with a relative
˚
occupancy of 73%. This Cr≡N bond distance of 1.585 A makes
˚
sense in the context of the Cr≡N bond distance of 1.575 A found in
[(Me3tacn)CrN(acac)]+,9 the only structurally characterized Cr(V)-
nitrido species bearing the tacn moiety as found in 9.
1868 | Dalton Trans., 2008, 1864–1871
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Thus, the re-refined model of 9 contains ∼27% of an impurity
that we propose to be a Cr(III) hydroxo complex, a reasonable
assignment considering the possibility of partial hydrolysis of
the Cr≡N group. The so-refined Cr–OH bond distance, 1.87(2)
over CaH2 and distilled under argon before use. Other chemicals
were obtained from commercial sources and used without further
purification.
˚
Synthesis of ligands and complexes
A, is also reasonable considering Cr–OH bond distances known
22
˚
from other Cr(III) complexes (mean of 1.92 A). The presence
1,4,7-Triazacyclononane, (tacn) was synthesized as described in
the literature.24 The lithium salts of the corresponding ligands
have been synthesized and characterized as reported in the
literature.15 Li2L1·2LiOH·H2O = {4-carboxymethyl-7-benzyl-
[1,4,7]triazonan-1-yl}-acetic acid lithium salt; Li2L2·LiOH·
2EtOH = {4-carboxymethyl-7-isopropyl-[1,4,7]triazonan-1-yl}-
acetic acid lithium salt; Li2L3·2LiOH·2EtOH·2H2O = {4-carbo-
xymethyl-7-(methoxyl-benzyl)-[1,4,7]triazonan-1-yl}-acetic acid
lithium salt.
of a Cr(III)-hydroxo species in ∼27% of the sample was largely
overlooked in the analysis of the spectroscopic features of 9,
mainly because we identified features showing the existence of
the Cr(V)-nitrido species instead of features associated with its
purity. Purity was assumed (incorrectly) because of the crystallinity
of the sample. In this connection, it should be mentioned that
the elemental analysis data for 9 (Found: C 44.57%, H 6.32%,
N 11.40%) are a good fit for [L3CrN]·3H2O (Calcd: C 44.72%,
H 6.46%, N 11.59%) but may equally well support the formula
[L3CrN0.73(OH)0.27]·3H2O suggested by the crystallographic anal-
ysis (Calcd: C 44.60%, H 6.45%, N 10.90%). In fact, the electronic
spectra of the “nitrido” complexes, which show d–d transitions at
∼560 nm, ∼430 nm and ∼330 nm, more closely resemble spectra
of the corresponding Cr(III) species than genuine Cr(V)-nitrido
[L1CrCl] (1). To a purple solution of Cr(THF)3Cl3 (374 mg,
1.0 mmol) in MeOH (20 cm3) was added Li2L1·2LiOH·H2O
(413 mg, 1.0 mmol) in MeOH (10 cm3) at room temperature.
The purple color of the solution changed to green and some green
precipitates appeared. Zn granules were added and the solution
became clear. It was refluxed for 8 h, during which time the color
of the solution changed from green, purple to dark purple in
the end. The reaction mixture was filtered and slow evaporation
of the filtrate at RT led to the formation of purple precipitates,
which were dried under vacuum. Yield: 52% (218 mg). Anal.
Calcd for the complex L1CrCl (C17H23N3O4CrCl, 420. 8 g mol−1)
C 48.52, H 5.51, N 9.99; Found C 48.75, H 5.88, N 10.22. ESI-MS
spectrometry (positive mode): m/z = 421 ([L1CrCl]+).
2
species that typically have the Ballhausen–Gray 2B2 → E1 (for a
C
4v ML4 = E species) transition at ∼420 nm as their lowest-energy
electronic transition.9,23
The corrected structure of 9 brings up important crystallo-
graphic issues that have been identified previously in cases of
“bond-stretch isomerism” where incorrect metal–ligand or metal–
metal bond distances have been assigned due to unresolved
crystallographic disorder in crystalline mixtures of structurally
similar compounds.16,20
[L2CrCl] (2) and [L3CrCl] (3). These compounds were pre-
pared similarly to 1.
Conclusions
[L2CrCl]·2H2O (2). Recrystallization of the compound in
H2O–MeOH (3 : 7) solution resulted in purple crystals after
one week. Yield: 56%. Anal. Calcd for the complex L2CrCl·2H2O
(C13H27N3O6CrCl, 408.8 g mol−1) C 38.19, H 6.66, N 10.28; Found
C 38.02, H 6.83, N 10.70. ESI-MS spectrometry (positive mode):
m/z = 395 ([L2CrCl + Na]+). IR (KBr cm−1): 3456 (s), 1634 (s),
1492 (m), 1453 (w), 1334 (s), 1303 (s), 1168 (w), 1082 (m), 1034
(w), 1005 (m), 922 (s), 852 (m), 827 (m), 721 (m).
We have therefore shown that new six-coordinate Cr(III)-chloro
species 1–3 react cleanly with sodium azide to form the cor-
responding azido species 4–6. Further photolysis of 4–6 forms
Cr(V)-nitrido species 7–9 that are partially hydrolyzed to their
corresponding Cr(III)-hydroxo counterparts. The hydroxo and ni-
trido species co-crystallize complicating the structural refinement
of 9. What at first appeared to be the longest Cr≡N triple bond
˚
distance yet observed (1.66 A) was found after re-refinement to be
[L3CrCl] (3). The compound was dried under vacuum. Yield:
63%. Anal. Calcd for the complex L3CrCl (C18H25N3O5CrCl,
450.8 g mol−1) C 47.95, H 5.59, N 9.32; Found C 48.22, H
5.30, N 9.07. (ESI-MS spectrometry (positive mode): m/z = 450
([L3CrCl]+). IR (KBr cm−1): 3455 (m), 1670 (s), 1517 (s), 1491 (m),
1456 (m), 1328 (s), 1295 (s), 1254 (m), 1245 (m), 1204 (w), 1180
(w), 1080 (m), 1031 (m), 958 (m), 922 (m), 830 (s), 729 (s), 642 (m),
565 (m).
an artifact of positional disorder of the Cr atoms of nitrido and
hydroxo species. The re-refined Cr≡N bond distance is estimated
˚
as 1.58 A, reasonably within the limits of known six-coordinate
Cr(V)-nitrido species. UV-vis data for 7–9 suggest that all three
nitrido species are partially hydrolyzed. These results emphasize
the care that must be taken in the characterization of compounds
that may co-crystallize with structurally similar analogs.
[L1CrN3] (4). To a solution of 1 (200 mg, 0.475 mmol) in
MeOH (20 cm3) was added NaN3 (309 mg, 4.75 mmol), and the
solution was refluxed for 6 h and stirred at RT overnight. Then
the reaction mixture was filtered. Slow evaporation of the filtrate
in the dark led to the formation of pink crystals suitable for X-ray
measurement. Yield: 46% (93 mg). Anal. Calcd for the complex
L1CrN3 (C17H23CrN6O4) C 47.77, H 5.42, N 19.66; Found C 48.13,
H 5.70, N 19.18. ESI-MS spectrometry (positive mode): m/z = 428
([L1CrN3 + H]+). IR (KBr cm−1): 3455 (m), 2064 (s), 1669 (s), 1492
(m), 1452 (w), 1319 (m), 1287 (m), 1077 (m), 1054 (m), 966 (w),
913 (m), 858 (w), 816 (w), 770 (m), 714 (m), 657 (w).
Experimental
Note
Although we have not encountered any problems in the present
study, it should be kept in mind that metal azides are potentially
explosive and should be handled carefully in small amounts and
with appropriate precaution.
The solvents diethyl ether, toluene, ethanol, dichloromethane,
chloroform and n-hexane were commercially available and dried
over molecular sieves before use; methanol and MeCN were dried
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1864–1871 | 1869
©

View Article Online
[L2CrN3] (5) and [L3CrN3] (6). These compounds were syn-
thesized similarly to 4.
(m), 1065 (m), 1030 (w), 1006 (w), 968 (m), 925 (m), 822 (s), 727
(m), 629 (w). giso = 1.98 (10 K).
[L2CrN3]·0.5H2O (5). Crystals were obtained from ether dif-
fusion into a MeOH–MeCN solution after two weeks. Yield: 30%.
Anal. Calcd for the complex L2CrN3·05H2O (C13H24CrN6O4.5,
388.38 g mol−1) C 40.17, H 6.18, N 21.62; Found C 39.87, H
6.06, N 21.23. ESI-MS spectrometry (positive mode): m/z = 402
([L2CrN3 + Na]+). IR (KBr cm−1): 3404 (s), 2067 (s), 1668 (s), 1493
(w), 1460 (w), 1384 (w), 1337 (s), 1303 (w), 1164 (w), 1082 (m),
1065 (m), 1032 (w), 1002 (w), 918 (m), 853 (m), 823 (m), 787 (w),
721 (w).
Physical measurements
Infrared spectra (400–4000 cm−1) of solid samples were recorded
on a Perkin-Elmer 2000 FT-IR/FT-NIR Spectrometer as KBr
disks. ESI-MS spectra were recorded on a Finnigan MAT 95
mass spectrometer. UV/Vis spectra of solutions were measured
on a Perkin-Elmer Lambda 19 spectrophotometer in the range
200–1000 nm. Temperature-dependent magnetic susceptibilities
of powdered samples were measured by using a SQUID magne-
tometer (Quantum Design) at 1.0 T (2.0–300 K). Corrections for
underlying diamagnetism were made by using tabulated Pascal
constant. X-band EPR spectra were recorded on a Bruker ESP
300E spectrometer equipped with a helium flow cryostat (Oxford
Instruments ESR 910).
[L3CrN3] (6). Purple crystals of 6 suitable for X-ray crystal-
lography were obtained from slow evaporation of MeOH solution
at room temperature after one week. Yield: 53%. Anal. Calcd
for the compound L3CrN3 (C18H25N6O5Cr, 457.4 g mol−1): C
47.26, H 5.51, N 18.37; Found C 47.60, H 5.44, N 18.69. ESI-
MS spectrometry (positive mode): m/z = 480 ([L3CrN3 + Na]+).
IR (KBr cm−1): 3441 (m), 2060 (s), 1684 (s), 1665 (s), 1511 (s), 1451
(m), 1329 (s), 1290 (s), 1254 (m), 1181 (m), 1081 (m), 1072 (m),
1056 (m), 1021 (s), 954 (w), 921 (m), 832 (m), 722 (m), 666 (w).
Low-temperature EPR, magnetization measurements were an-
alyzed on the basis of a spin-Hamiltonian description of the
electronic ground-state spin multiplet
2
2
2
He = D[Sz − S(S + 1)/3 + (E/D)(Sx − Sy )] + lBB gS
[L3CrN]·3H2O (9). A solid sample of 6 (70 mg, 0.153 mmol)
was dissolved in 30 cm3 of MeCN–H2O = (1 : 1) and the
resulting purple solution was photolyzed at ambient temperature
with Rayonet Photochemical Reactor (RPR-100) equipped with
419 nm tubes. The photolysis process was followed optically by
changes in the color of the species from dark purple to light
pink for 10 h at room temperature. An EPR measurement of the
resulting solution at 10 K shows a typical Cr(V) signal centered at
giso = 1.98. At the same time, small amounts of photolyzed solution
was taken out and allowed to evaporate under vacuum to yield a
light pink solid. IR measurements show that the azide stretch at
2060 cm−1 disappears completely and a new peak due to Cr≡N
stretch appears at 971 cm−1. Slow evaporation of the solution at
room temperature followed by cooling the concentrated solution
where S = 3/2 or S = 1/2 is the spin of the corresponding system
and D and E/D are the axial and rhombic zero-field parameters
for S > 1/2.
X-Ray crystallography
All the crystals were covered with perfluorinated polyether and
mounted on glass fibers. Graphite monochromated Mo Ka radi-
˚
ation (k = 0.71073 A) was used throughout. Intensity data were
collected on a Siemens SMART CCD-detector system equipped
with a cryogenic nitrogen cold stream at 100 K. Data collection was
performed by a hemisphere run taking frames at 0.30 ◦ in x. The
data were corrected for Lorentz and polarization effects. A semi-
empirical absorption correction using the program SADABS25
was performed for the compounds. The Siemens SHELXTL26,27
software package was used for solution, refinement, and artwork
of the structures. All the structures were solved and refined by
direct methods and difference Fourier techniques. Neutral atom
scattering factors were used. All non-hydrogen atoms were refined
anisotropically. All hydrogen atoms were placed at calculated
positions and refined as riding atoms with isotropic displacement
parameters.
◦
at 4 C resulted in the formation of pink crystals suitable for X-
ray crystallography. Anal. Calcd for the complex L3CrN·3H2O
(C18H31CrN4O8, 483.5 g mol−1) C 44.72, H 6.46, N 11.59; Found
C 44.57, H 6.32, N 11.40. ESI-MS spectrometry (positive mode):
m/z = 452 ([L3CrN + Na]+). Yield: 78% (51 mg). IR (KBr cm−1):
3430 (m), 1655 (s), 1512 (s), 1364 (m), 1334 (m), 1300 (m), 1251
(m), 1244 (m), 1179 (w), 1086 (w), 1073 (w), 1034 (m), 971 (m),
922 (m), 830 (m), 726 (m), 635 (w).
L1CrN·3H2O (7) and L2CrN·2H2O (8). These compounds
were synthesized in a similar way by photolysis of the correspond-
ing azido-Cr(III) precursors in MeCN–H2O (1 : 1) at 419 nm until
the purple color of the solution was bleached and a light pink
solution was obtained after 10 h. Anal. calcd for L1CrN·3H2O
(C17H29CrN4O7, 453.4 g mol−1) C 45.03, H 6.45, N 12.36; Found
C 45.34, H 6.02, N 12.14. ESI-MS spectrometry (positive mode):
m/z = 423 ([L1CrN + Na]+); IR (KBr cm−1): 3428 (s), 1635 (s),
1496 (m), 1461 (w), 1387 (m), 1337 (m), 1297 (m), 1073 (m), 1036
(w), 1010 (m), 971 (m), 925 (m), 858 (w), 838 (w), 816 (w), 767 (m),
710 (m). giso = 1.98 (10 K).
Computations
Calculations were performed using the ORCA program.28 Geome-
try optimization utilized a truncated model ligand (Scheme 2) with
all atomic coordinates taken from the crystal structure of 9 before
re-refinement of the disordered Cr atom. The BP86 functional29–31
was used in the geometry optimization, with a triple-f basis set.32
Acknowledgements
Anal. Calcd, for L2CrN·2H2O (C13H27N4O6Cr, 387.4 g mol−1),
C 40.31, H 7.03, N 14.46 ESI-MS spectrometry (positive mode):
m/z = 352 ([L1CrN + H]+); IR (KBr cm−1): 3421 (s), 1653 (s),
1492 (w), 1458 (w), 1384 (m), 1336 (m), 1291 (m), 1125 (w), 1094
We thank Karl Wieghardt for his guidance on this project.
Financial support of this work by the Fonds der Chemischen
Industrie is acknowledged.
1870 | Dalton Trans., 2008, 1864–1871
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
16 G. Parkin, Chem. Rev., 1993, 93, 887, and references therein.
17 J. Bendix, R. J. Deeth, T. Weyhermu¨ller, E. Bill and K. Wieghardt,
Inorg. Chem., 2000, 39, 930.
18 J. Bendix, T. Birk and T. Weyhermu¨ller, Dalton Trans., 2005, 2737.
19 J. F. Berry, E. Bill, E. Bothe, S. D. George, B. Mienert, F. Neese and K.
Wieghardt, Science, 2006, 312, 1937.
20 J. F. Berry, F. A. Cotton, T. B. Lu, C. A. Murillo, B. K. Roberts and
X. P. Wang, J. Am. Chem. Soc., 2004, 126, 7082.
References
1 K. B. Wiberg, in Oxidation in Organic Chemistry, ed. K. B. Wiberg,
Academic Press, New York, 1965, Part A, p. 69.
2 W. A. Nugent and J. M. Mayer, Metal–Ligand Multiple Bonds, John
Wiley & Sons, New York, 1988.
3 S. I. Arshankow and A. L. Poznjak, Z. Anorg. Allg. Chem., 1981, 481,
201.
4 J. T. Groves, T. Takahashi and W. M. Butler, Inorg. Chem., 1983, 22,
884.
5 C. M. Che, J. X. Ma, W. T. Wong, T. F. Lai and C. K. Poon, Inorg.
Chem., 1988, 27, 2547.
21 M. Nippe and J. F. Berry, J. Am. Chem. Soc., 2007, 129, 12684.
22 M. Ardon and A. Bino, Inorg. Chem., 1985, 24, 1343.
23 C. J. Ballhausen and H. B. Gray, Inorg. Chem., 1962, 1, 111.
24 K. Wieghardt, W. Schmidt, B. Nuber and J. Weiss, Chem. Ber./Recl.,
1979, 112, 2220.
25 G. M. Sheldrick, SADABS V 2.03: Area Detector Absorption Correc-
tion, University of Go¨ttingen, Go¨ttingen, Germany, 1999.
26 G. M. Sheldrick, SHELXS-97, Program for Crystal Structure Solution,
University of Go¨ttingen, Go¨ttingen, Germany, 1997.
27 G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Refine-
ment, University of Go¨ttingen, Go¨ttingen, Germany, 1997.
28 F. Neese, ORCA - an ab initio, DFT, and Semiempirical Electronic
Structure Package, V 2.6.0, Universita¨t Bonn, Bonn, Germany,
2003.
6 H. Grunewald and H. Homborg, Z. Anorg. Allg. Chem., 1992, 608, 81.
7 J. Bendix, J. Am. Chem. Soc., 2003, 125, 13348.
8 T. Birk and J. Bendix, Inorg. Chem., 2003, 42, 7608.
9 A. Niemann, U. Bossek, G. Haselhorst, K. Wieghardt and B. Nuber,
Inorg. Chem., 1996, 35, 906.
10 K. Meyer, J. Bendix, E. Bill, T. Weyhermu¨ller and K. Wieghardt, Inorg.
Chem., 1998, 37, 5180.
11 G. D. Fallon, G. A. McLachlan, B. Moubaraki, K. S. Murray, L. Obrien
and L. Spiccia, J. Chem. Soc., Dalton Trans., 1997, 2765.
12 L. F. Lindoy, The Chemistry of Macrocyclic Ligands, Cambridge
University Press, Cambridge, 1989.
29 A. D. Becke, Phys. Rev. A, 1988, 38, 3098.
13 A. McAuley, S. Subramanian, M. J. Zaworotko and R. Atencio, Inorg.
30 J. P. Perdew, Phys. Rev. B, 1986, 33, 8822.
Chem., 1998, 37, 4607.
31 J. P. Perdew, Phys. Rev. B, 1986, 34, 7406.
14 A. C. Warden, L. Spiccia, M. T. W. Hearn, J. F. Boas and J. R. Pilbrow,
Dalton Trans., 2005, 1804.
15 Y. F. Song, J. F. Berry, E. Bill, E. Bothe, T. Weyhermu¨ller and K.
Wieghardt, Inorg. Chem., 2007, 46, 2208.
32 A. Schafer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97,
2571.
33 J. Bendix, K. Meyer, T. Weyhermu¨ller, E. Bill, N. Metzler-Nolte and K.
Wieghardt, Inorg. Chem., 1998, 37, 1767.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1864–1871 | 1871
©