High-field EPR and magnetic susceptibility studies on binuclear and tetranuclear copper trifluoroacetate complexes. X-ray structure determination of three tetranuclear quinoline adducts of copper(II) trifluoroacetate

Article abstract of DOI:10.1021/ja902695y

Magnetic properties and high-field EPR spectra of three previously unknown tetranuclear quinoline adducts of copper(II) trifluoroacetate were studied, and their X-ray structures were determined. Two green complexes containing a μ₄-oxo bridge, C

Full text of DOI:10.1021/ja902695y

Published on Web 07/06/2009
High-Field EPR and Magnetic Susceptibility Studies on
Binuclear and Tetranuclear Copper Trifluoroacetate
Complexes. X-ray Structure Determination of Three
Tetranuclear Quinoline Adducts of Copper(II) Trifluoroacetate
,†
‡
‡
§
Andrew Ozarowski,* Iwona B. Szyma n´ ska, Tadeusz Muzioł, and Julia Jezierska
National High Magnetic Field Laboratory, Florida State UniVersity, 1800 East Paul Dirac
DriVe, Tallahassee, Florida 32310, Department of Chemistry, Nicolaus Copernicus UniVersity,
ul. Gagarina 7, Toru n´ 87100, Poland, and Faculty of Chemistry, Wrocław UniVersity,
ul. F. Joliot-Curie 14, Wrocław 50383, Poland
Received April 3, 2009; E-mail: ozarowsk@magnet.fsu.edu
Abstract: Magnetic properties and high-field EPR spectra of three previously unknown tetranuclear quinoline
adducts of copper(II) trifluoroacetate were studied, and their X-ray structures were determined. Two green
complexes containing a µ
1
2
4
-oxo bridge, Cu
5.278(3), b ) 23.227(5), c ) 34.895(7) Å) and Cu
1.933(4), b ) 11.176(2), c ) 23.927(5) Å, ꢀ ) 97.41(3)°) are ferromagnetic, whereas the blue complex
4
O(CF
3
COO)
6
(quin)
4
·(C
6
H
5
CH
3
)
0.6 (orthorhombic, Pbca, a )
4
O(CF
3
COO)
6
(quin)
4
·(C
6 6
H )
0.8 (monoclinic, P2/c, a )
2-
+
[
9
Cu
4
(OH)
7.41(3)°), formed in humid air from the solid dimeric [Cu(CF
-oxo tetramers. High-field EPR spectra allowed
2
(CF
3
COO)
8
(quin)
2
]
(quinH )
2
(monoclinic, a ) 21.933(4), b ) 11.176(2), c ) 23.927(5) Å, ꢀ )
3
COO) ·(quin)] , is antiferromagnetic, as is
2
2
the tetranuclear blue product formed in humid air of the µ
4
determination of the spin Hamiltonian parameters for the spin quintet state (S ) 2) in the ferromagnetic
complexes, which facilitated accurate interpretation of their magnetic susceptibility data. “Broken symmetry”
DFT calculations were performed to estimate the exchange integrals in all three tetranuclear complexes,
showing surprisingly good agreement with experimental results. Negative sign of the zero-field splitting
parameter D in two binuclear complexes, [Cu(CF
3
COO)
2
·CH
3
CN]
2
and [Cu(CF
3
COO)
2
2
·(quin)] , was found
from single-crystal high-field EPR spectra, confirming recent results for nonhalogenated dimeric copper
carboxylates.
6
2
Introduction
[Cu(CF COO) ·(quin)] (quin ) quinoline) are rare exceptions.
3
2
Copper(I) and copper(II) perfluorocarboxylates are volatile,
which makes them potentially useful in the chemical vapor
deposition technique to produce thin metallic copper layers,
Binuclear and polynuclear copper carboxylates have been
subject to extensive experimental and theoretical studies ever
since Bleaney and Bowers discovered intramolecular antifer-
romagnetic interactions in binuclear copper acetate monohy-
7
and from that viewpoint the determination of nuclearity of such
complexes is very important. Magnetic susceptibility measure-
ments and electron paramagnetic resonance are methods of
choice to accomplish this task. In this way, these complexes
are on the crossroads of pure and applied chemistry and offer
an opportunity of investigating the metal-metal interactions in
systems substantially different from the better-known nonha-
1
drate. Factors affecting the magnitude and character of the
exchange interactions in bridged binuclear transition metal
2
,3
complexes have been of particular interest.
Dimeric “paddlewheel” structures are very conspicuous
3
among copper(II) complexes of formic acid and its homologues,
8
while such arrangements are much less numerous in the case
of the perfluorinated carboxylic acids, which tend to form either
logenated copper carboxylates. One of us has recently inves-
tigated by high-field EPR the zero-field splitting in copper
acetate monohydrate and its pyrazinate adduct to find that the
zero-field splitting parameter D is negative. |D| in [Cu(CF3-
4
monomeric copper complexes or more extended chains.
5
Dimeric paddlewheel adducts [Cu(CF
3
COO)
2
·(CH
3
CN)]
2
and
2 2
COO) (quin)] has been reported to be substantially larger than
6
†
in simple copper carboxylates, and this research was undertaken
in part to determine whether D carries a negative sign in the
trifluoroacetate copper complexes as well.
Florida State University.
Nicolaus Copernicus University.
Wrocław University.
‡
§
(
(
1) Bleaney, B.; Bowers, K. D. Proc. R. Soc. A 1952, 214, 451–465.
In the course of this work we encountered the surprisingly
rich system Cu(II)-trifluoroacetate-quinoline. In addition to the
2) Hay, P. J.; Thibeault, J. C.; Hoffmann, R. J. Am. Chem. Soc. 1975,
9
7, 4884–4899.
(
(
3) Melnik, M. Coord. Chem. ReV. 1981, 36, 1–44.
4) Cotton, F. A.; Dikarev, E. V.; Petrukhina, M. A. Inorg. Chem. 2000,
(6) Moreland, J. A.; Doedens, R. J. J. Am. Chem. Soc. 1975, 97, 508–
513.
3
9, 6072–6079.
(
5) Karpova, E. V.; Boltalin, A. I.; Zakharov, M. A.; Sorokina, N. I.;
(7) Grodzicki, A.; Lakomska, I.; Piszczek, P.; Szymanska, I.; Szlyk, E.
Coord. Chem. ReV. 2005, 249, 2232–2258.
Korenev, Yu. M.; Troyanov, S. I. Z. Anorg. Allg. Chem. 1998, 624,
7
41–744.
(8) Ozarowski, A. Inorg. Chem. 2008, 47, 9760–9762.
10.1021/ja902695y CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 10279–10292 9 10279

A R T I C L E S
Ozarowski et al.
6
dimer mentioned above, two green ferromagnetic tetranuclear
product that will be called 2a here was described previously, and
complexes Cu
4
O(CF COO)
3
6
(quin)
4
· (C
6
H
5
CH
3
)
0.6 and Cu
4
O-
3 2 9 7 2
the formula Cu(CF COO) (C H N)(H O) was proposed on the basis
of chemical analysis only. In our case, the yellow-green crystals
of 2 converted to blue on humid air, seemingly without crystal
destruction. However, the product turned out not to consist of single
crystals. Blue crystalline product of X-ray quality was obtained by
exposing solution of dimer 2 in benzene to humid air and an
unprecedented tetranuclear structure was revealed by X-ray dif-
fractometry (see below). IR spectra confirmed that the two blue
products are the same compound. Warming the blue powder 2a to
(
CF COO) (quin)
3
6
4
·(C H )0.8 were synthesized, and their struc-
6 6
tures were determined. A blue product formed from the
tetranuclear complexes on exposure to humid air was also found
to be tetranuclear but antiferromagnetic. Also, the blue antifer-
2
-
+
romagnetic complex [Cu
4
(OH) COO)
2
(CF
3
8 2 2
(quin) ] (quinH )
formed from [Cu(CF COO)
3
2
(quin)] on exposure to humid air
2
was found to be tetrameric, and its X-ray structure was resolved.
∼
80 °C converts it back to the yellow-green 2. Apparently, 2 and
Experimental Section
2
a convert to each other in the process
The X-band EPR spectra were recorded on a Bruker ElexSys
spectrometer (Bruker, Germany). High-frequency EPR spectra at
temperatures ranging from ca. 3 to 290 K were recorded on a home-
2
[Cu(CF COO) (quin)] + 2H O T
3 2 2 2
9
2-
+
built spectrometer at the EMR facility of NHMFL. The instrument
[Cu (OH) (CF COO) (quin) ] (quinH )
4
2
3
8
2
2
was a transmission-type device in which microwaves are propagated
in cylindrical lightpipes. The microwaves were generated by a
phase-locked Virginia Diodes source generating a frequency of 13
6
The Moreland’s formula Cu(CF
3
COO)
2
(C
9
H
7
N)(H
2
O) contains
two more water molecules per tetrameric unit than ours for 2a
molar mass 1711.0) and the two formulas are difficult to distinguish
(
1 GHz and producing its harmonics of which the fourth, eighth,
6th, 24th, and 32nd were available. A superconducting magnet
Oxford Instruments) capable of reaching a field of 17 T was
employed.
Magnetic susceptibility data of powdered samples were measured
(
1
by chemical analysis.
(
Analysis of 2a. Cu 14.8, C 36.7, H 2.69, N 3.45; calcd for
Cu
Cu
COO)
4
C
52
F
24
H
32
N
3
4
O
18 Cu 14.86, C 36.50, H 1.88 N 3.27.
COO) (quin) ·(C CH 0.6 (3) and Cu
·(C 0.8 (4). These substances can be obtained from
4
O(CF
6
4
6
H
5
3
)
4 3
O(CF -
with a SQUID magnetometer (Quantum Design MPMSXL-5) over
the temperature range 1.8-300 K at the magnetic induction of 0.5
T. Corrections for the sample holders were applied. Diamagnetic
corrections for the molecules were determined from Pascal’s
6
(quin)
4
6 6
H )
the crude green powder (above) or from the blue product 2a.
Approximately 1 g of the substrate was dissolved by boiling in 30
mL of benzene or toluene, respectively. The solution was left to
cool slowly in a tightly closed flask causing precipitation of small
green crystals of 3 or 4 within hours. Obtaining X-ray quality
crystals was difficult. IR bands for 3 (cm ): 3422, 3093, 2948,
853,1686, 1599, 1559, 1513, 1442, 1406, 1378, 1316, 1199, 1146,
058, 960, 846, 810, 795, 781, 727, 638, 613, 522, 497, 468. IR
1
0
constants.
Single-Crystal Structure Determination. The single-crystal
diffraction data for 2a, 3, and 4 were collected at room temperature
on an Oxford Diffraction KM4 CCD diffractometer employing Mo
KR radiation with λ ) 0.71073 Å. The reflections were measured
with ω-2θ method and numerical absorption corrections were
-1
2
1
-1
1
1
bands for 4 (cm ): 3422, 3085, 2954, 2859,1686, 1599, 1561, 1513,
1443, 1406, 1379, 1316, 1199, 1146, 1058, 960, 845, 810, 795,
applied. All figures have been prepared in ORTEP and DIA-
1
2
MOND v. 2.1e programs.
Synthesis. A blue solution of Cu(CF
reacting CF COOH with CuO, was evaporated to dryness, and the
resulting green solid was used in subsequent preparations.
Cu(CF COO) ·(CH CN)] (1). Copper trifluoroacetate was
dissolved in CH CN to form a saturated solution at ca. 60 °C in a
7
82, 727, 638, 613, 522, 497, 468.
Analysis of 3 [Cu O(CF COO) (quin)
C 41.0, H 2.7, N 3.8; calcd for Cu
520.2) Cu 16.72, C 41.24, H 2.17, N 3.68.
Analysis for 4. Cu 16.7, C 40.4, H 2.8, N 3.98; calcd for
Cu 13 (MW ) 1527.4) Cu 16.64, C 41.52, H 2.16,
N 3.67:
3 2
COO) , prepared by
4
3
6
4
6 5 3
]·0.6C H CH . Cu 16.7,
3
4
52.2 32.8 4 18
C H N F O13 (MW )
1
[
3
2
3
2
3
4 52.8 32.8 4 18
C H N F O
closed vial. Slow temperature lowering resulted in formation within
hours of very large (up to 10 mm × 3 mm × 2 mm) green,
transparent crystals. The substance is very hygroscopic and flows
when exposed to humid air. The identity of the product, which was
Some samples of 3 and 4 were found by EPR to be contaminated
by 2, but very pure crystalline samples of all three substances were
also prepared. Both 3 and 4 convert on humid air to a blue solid
product 3a, which is different from 2a. Heating 3a does not convert
it back to 3 nor to 2, but 3 or 4 can be obtained by boiling 3a in
toluene or benzene, respectively. Analytical data for 3a: Cu 16.1,
C 38.8, H 2.9, N 3.9.
CCDC 733624, 733605, and 733606 contain the supplementary
crystallographic data for 2a, 3, and 4, respectively. These data can
be obtained free of charge from the Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Note on the Analysis. Fluorine-containing compounds pose
difficulties when using standard instrumentation for the C, H, N
analysis, which in our case were augmented by copper trifluo-
roacetate volatility and are reflected in our poor results for H.
Problems were also encountered in AAS analysis and for that reason
5
described previously, was confirmed by the X-ray determination
of its crystal lattice parameters.
[
Cu(CF
3
COO)
2
·(quin)]
2
(2). The method described in ref 6 was
followed. Violet powder Cu(CF
3
COO) (quin) , obtained by reacting
2
2
copper trifluoroacetate with quinoline in ethanolic solution, was
heated to ca. 80 °C in a Petri dish until it turned green. The crude
green product was dissolved in 99.8% benzene or in toluene and
filtered. The commercial solvents were not dried, which may have
aided in obtaining some of the compounds described below.
Evaporation of a solution at elevated temperature, close to the
solvent boiling point and under protection from humidity, resulted
in formation of large yellow-green X-ray quality crystals of 2. IR
bands: 3422, 3094, 2955, 2853, 1704, 1599, 1565, 1512, 1441,
13
1
404, 1379, 1315, 1198, 1146, 1056, 958, 846, 808, 795, 780, 728,
13
copper was also determined iodometrically.
-1
6
13, 522, 497, 475 cm . The product identity was confirmed by
6
X-ray determination of its lattice constants.
Results and Discussion
2-
+
[
Cu
4 2 3 8 2 2
(OH) (CF COO) (quin) ] (quinH ) (2a). Compound 2
turns blue on prolonged exposure to humid air. This blue solid
IR Spectra. Asymmetric and symmetric stretching vibrations
of the carboxylic group were detected at: νas(COO) ) 1686 cm
s
for 3 and 4 and at 1704 cm for 2, while ν (COO) appeared at
-1
(
9) Hassan, A. K.; Pardi, L. A.; Krzystek, J.; Sienkiewicz, A.; Goy, P.;
-1
Rohrer, M.; Brunel, L.-C. J. Magn. Reson. 2000, 142, 300–312.
(
(
10) O’Connor, C. J. Prog. Inorg. Chem. 1982, 29, 203–283.
11) CrysAlis RED and CrysAlis CCD; Oxford Diffraction: Abingdon,
Oxfordshire, England, 2000.
(13) Campbell, A. D.; MacDonald, A. M. G. Anal. Chim. Acta 1962, 26,
275–280. Hall, W. T., Williams, R. S. In The Chemical and
Metallographic Examination of Iron, Steel and Brass; McGraw-Hill:
New York, 1921.
(
12) Brandenburg, K. DIAMOND, Release 2.1e; Crystal Impact GbR: Bonn,
Germany, 2001. Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
1
0280 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
-
1
1
s
441-1443 cm . The difference ∆ν ) νas(COO) - ν (COO)
-1
-1
for 3 and 4 is 244 cm but is 263 cm for 2. These values are
in accord with the bridging carboxylate coordination in 2, 3,
and 4. The νring and δring band shifts, typical for coordinated
1
4
-1
15
quinoline, were observed over the range 600-1600 cm . In
-
1
the spectra of 3 and 4 a sharp absorption band at 638 cm was
detected. This band is missing from the spectra of the free
quinoline, trifluoroacetates, and also of complex 2. Therefore
it was assigned to the vibration of the Cu
4
O core. In the case of
1
6
compounds [Cu
O)(OSR ] (R ) Et, Bu), the Cu
66-583 cm , whereas they appeared at 581-592 cm in
4
(µ
4
-O)(bahped)
2
](ClO
4
)
2
and [Cu
4
Cl
6
(µ
4
-
n
17
2
)
4
4
O bands were observed at
-1
-1
5
18
copper 2,6-bis(morpholinomethyl)-4-methylphenol complexes.
Blue Products 2a and 3a Compared to Green Complexes
2
-4. Although the carboxylate ∆ν of 254 cm- in 3a is altered
1
compared to tetramers 3 and 4, it is still within the bridging
19
4
carboxylate range. The µ -O bridge seems to be retained in 3a.
There is some IR evidence that in 3 and 4 quinoline has been
oxidized on exposure to humid air to form quinoline N-oxide
that interacts with copper via the oxygen atom from the highly
polarized NfO group. The oxidation may not have occurred
for all quinoline molecules, because the new bands ν(CdN) )
Figure 1. Molecular structure of 3. The copper, oxygen, nitrogen, carbon,
and fluorine atoms are drawn in magenta, red, blue, gray, and green,
respectively. Hydrogen atoms are omitted and only one of the disordered
positions of quinoline and of each disordered fluorine atom is shown.
Thermal ellipsoids are drawn at 20% for the image clarity. The molecular
structure of 4 is shown in Figure S7 in Supporting Information.
1
590, ν(N-O) ) 1272, ν(N-O) ) 1244 add to the old ones
rather than replacing them, and no effect on EPR parameters
of replacing N by O was observed for the blue product 3a (see
below). The intensity increase of the bands at 1145 and 1163
Structure Description. Important bond lengths and angles
-1
cm may be attributed to the appearance of the δ(N-O)
for 3 and 4 are given in Table 2. [Cu
4 3 6
O(CF COO) -
vibration. The off-plane C-H bands are seen at 810 and 727
(
quin) ]·0.6C CH (3) crystallizes in the orthorhombic Pbca
4
6
H
5
3
-1
cm . The changes in IR described here are accompanied by
drastic changes in the magnetic properties and EPR spectra.
space group. The asymmetric unit contains 1 tetranuclear
molecule and 0.6 molecules of toluene.
-
1
Blue species 2a exhibits no µ
4
-O band. ∆ν of 250 cm is
2-
The O anion O5 at the molecule center is surrounded by
-
1
within the range of bridging carboxylate. The 1404 cm band
splits into two bands at 1400 and 1412 cm , and the 785 cm
four copper(II) ions forming a distorted tetrahedron with the
Cu-Cu distances varying from 3.069 to 3.376 Å. The coordina-
tion sphere of each copper atom adopts a distorted square
pyramid geometry owing to the Jahn-Teller effect. The copper
coordination sphere consists of three oxygen atoms of carboxy-
-1
-1
band intensity is reduced compared to that of 3a.
X-ray Structure of [Cu
4
O(CF
3
COO)
6
(quin)
4
]·0.6C
6 5 3
H CH
(
3, Figure 1). The details of data collection are given in Table
1
. The crystal dimensions were 0.10 mm × 0.06 mm × 0.62
2-
late anions, one central O anion, and one nitrogen atom of
mm. The structure was solved by direct methods and refined
with a full-matrix least-squares procedure on F using the
SHELX-97 program package. Because of the high dynamics
exhibited by the CF groups and one of the quinoline molecules,
the quinoline molecule. The nitrogen atom in the base of each
2
square pyramid is bound in trans position versus the central
2
0
2-
O
anion. Similarly, two carboxylate oxygen atoms are in trans
3
position versus each other. The coordination spheres are
completed by the third carboxylate oxygen atom coordinated
in the axial position, at distances of 2.257(12)-2.372(11) Å,
which are longer by 0.3-0.4 Å than those of the carboxylate
oxygen atoms found in the basal plane. The Cu-O(carboxylate)
bond lengths in the basal plane (1.929(12)-1.992(11) Å) are
similar to the Cu-O5 bond lengths (1.949(8) and 1.969(9)
Å). Nitrogen is bound at a slightly longer distance
geometrical and/or thermal restraints were applied for these parts
of the structure. Similarly, such restraints were applied for the
partially occupied toluene sites. Hydrogen atoms were added
according to geometry and not refined. They were omitted for
toluene molecule as it exhibits disorder and high temperature
factors leading to a low-quality electron density maps in this
region. This data set is of the lowest quality of the three reported
here. As a consequence, the goodness-of-fit and R parameters
are slightly higher than in 4 and 2a.
(
2.009(17)-2.035(14) Å).
All carboxylates are involved in bridging interactions.
However, there are two types of carboxylate bridges. In the first
one a carboxylate joins an axial position of one copper atom
with an equatorial position of another copper atom. The other
type of bridge is symmetric and joins equatorial positions of
two copper atoms. There are four bridges of the former and
two of the latter type in the tetranuclear molecule. The longest
intermetallic distances are found for the Cu atoms that are joined
by the symmetric carboxylate groups. These intermetallic
distances as well as the torsion angles M-O-C-O/O-C-O-M
(-9.93°/34.87°) indicate that copper atoms are bound in a syn/
syn arrangement (ideal values 0°/0°). The distances C-O(carbox-
ylate) range from 1.20(2) to 1.25(2) Å, and this bond length
depends neither on the carboxylate coordination mode nor on
(
14) Nakamoto, K. In Infrared and Raman Spectra of Inorganic and
Coordination Compounds, part B; John Wiley & Sons: New York,
1
997; pp 59-62; 276-282.
15) Ozel, A. E.; Buyukmurat, Y.; Akyuz, S. J. Mol. Struct. 2001, 565-
66, 455–462.
16) Bera, M.; Wong, W. T.; Aromi, G.; Ribas, J.; Ray, D. Inorg. Chem.
(
(
(
(
(
(
5
2
004, 43, 4787–4789.
17) Guy, J. T., Jr.; Cooper, J. C.; Gilardi, R. D.; Flippen-Anderson, J. L.;
George, C. F., Jr. Inorg. Chem. 1988, 27, 635–638.
18) Teipel, S.; Griesar, K.; Haase, W.; Krebs, B. Inorg. Chem. 1994, 33,
4
56–464.
19) Kalyanasundaram, R.; Navaneetham, N. S.; Soundararajan, S. Proc.
Indian Acad. Sci. 1985, 95, 235–242.
20) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122. Sheldrick,
G. M. SHELXS97, SHELXL97, and CIFTAB; University of G o¨ ttingen:
Germany, 1997.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10281

A R T I C L E S
Ozarowski et al.
Table 1. Details of Crystallographic Data Collection and Structure Refinement for 3 and 4
compound
toluene solvate (3)
benzene solvate (4)
32.8Cu
formula
C
52.2
H
32.8Cu F N O
4 18 4 13
52.8
C H
4 18 4 13
F N O
formula weight
crystal system
space group
1520.2
orthorhombic
Pbca (No. 61)
1527.4
monoclinic
P2/c (No. 13)
unit cell dimensions
15.278(3)
23.227(5)
34.895(7)
90
12383(4)
8
1.630
a [Å]
b [Å]
c [Å]
21.933(4)
11.176(2)
23.927(5)
97.41(3)
5816(2)
4
1.709
1.564
2971
ꢀ
[3deg]
V [Å ]
Z
D
3
3
calc [Mg/m ] [g/cm ]
absorption coefficient [mm-1]
F(000)
1.471
5808
crystal size [mm]
θ range for data collection [deg]
index ranges
0.10 × 0.06 × 0.62
2.11-18.85
-13 < h < 13
0.51 × 0.22 × 0.08
2.37-20.81
-21 < h < 21
-8 < k < 11
-
21 < k < 18
31 < l < 31
-
-23 < l < 23
reflections collected/independent reflections
transmission maximum/minimum
refinement method
41075/4861 [R(int) ) 0.1223]
23482/6077 [R(int) ) 0.0704]
0.9308/0.6107
0.8904/0.4996
full-matrix least-squares on F2
2
goodness-of-fit on F
1.266
1.075
a
final R indices [I > 2σ(I)]
R1 ) 0.1049, wR2 ) 0.2294
R1 ) 0.1098, wR2 ) 0.2330
0.485/-0.333
R1 ) 0.0828, wR2 ) 0.1981
R1 ) 0.1071, wR2 ) 0.2195
0.683/-0.905
a
R indices (all data)
largest diff peak and hole [e A-3]
a
2
2)2/Σ(w|F | ) ]1/2.
2 2
R1 ) Σ|F
o
| - |F
c
o
||/Σ|F |. wR2 ) [Σw(F
o
- F
c
o
its contacts to copper. Disorder of one quinoline molecule and
of all but one CF groups is observed. Assumption of the
alternative orientations for quinoline and the CF groups led to
a significant model improvement. There are numerous H-bonds
between the quinoline H atoms and fluorine atoms of CF , which
stabilize both the disordered orientations of CF groups and of
the quinoline molecule. In the latter case both orientations are
related approximately by a rotation about the N40-Cu4-N50
angle bisector. The distance Cu-N for the alternative disordered
nitrogen atom is the longest one, while the axial bond elongation
is the shortest. This additional position of nitrogen atom leads
to a very unfavorable geometry of the coordination polyhedron,
as the bond angles formed by this copper atom are far from
molecule as it exhibits partial occupancy and high temperature
factors leading to low quality of electron density maps in this
region.
3
3
4 3 6 4 6 6
Structure Description. The [Cu O(CF COO) (quin) ]·0.8C H
3
complex crystallizes in the monoclinic centrosymmetric P2/c
space group. The asymmetric unit contains one tetranuclear
molecule and a benzene molecule that was refined with 0.8
occupancy and positioned around a 2-fold axis. Two atoms of
benzene molecule (C121 and C123) are positioned strictly on
the 2-fold axis in special positions. C122 and C124 atoms and
their images generated by the 2-fold axis (-x, y, -z + 1/2)
complete the benzene molecule. The molecular structure is
essentially the same as that of the toluene solvate. The Cu-
O(carboxylate) bond lengths in the basal plane (1.949(9)-1.992(7)
Å)aresimilartotheCu-O5bondlengthsin3(1.949(7)-1.966(6)
Å). The nitrogen atom of the ordered quinoline molecules is
bound at a slightly longer distance (2.022(11)-2.034(12) Å).
The axial oxygen atoms were found at distances 2.297(7)-2.389-
(11) Å.
3
9
0° or 180° expected for a square pyramid. The ordered aromatic
quinoline rings are planar with rmsd 0.02-0.04 Å, whereas the
rmsds for the disordered one are 0.069 and 0.097 Å.
Cu1, Cu2, and Cu3 atoms are removed by approximately 0.05
Å (0.056-0.057 Å) from their basal planes toward the axial
oxygen atom. The disordered quinoline molecule adopting two
discrete positions defines two distinct basal planes that differ
only by one nitrogen atom (N40 or N50). The Cu4 atom leans
out by -0.016 and 0.191 Å from the basal planes involving
N40 and N50 atoms, respectively. In the first case Cu moves in
the direction opposite to the axial O71 oxygen atom, contrary
to the shifts observed for Cu1, Cu2, and Cu3 atoms. In the
second case the Cu4 atom shifts out of the N50 basal plane by
The most significant differences between structures of 3 and
4 concern the geometry around the Cu4 atom, which is
coordinated by the disordered quinoline molecules. In 4 we
observe additional electron density picks lying out of the N40
ring plane, whereas in 3 both orientations are related by a
bisector of the N40-Cu4-N50 angle. This time, the alternative
orientation (N50) reveals rocking by approximately 1 Å and
tilting by 35.9° versus the N40 ring, and it is stabilized by a
very short intramolecular H-bond to F85. The Cu-N bond
lengths found for the disordered quinoline molecule present two
extreme values: Cu-N40 is the longest (2.13 Å), whereas this
distance in the alternative orientation is in the usual range found
in this structure (2.01 Å) and falls within the typical range of
the Cu-O bond lengths.
0
.191 Å toward O71, as was observed for other Cu atoms.
However, this basal plane is significantly distorted as it shows
rmsd equal to 0.269 Å, whereas in other cases it does not exceed
0
.08 Å. The crystal packing and the hydrogen bond system is
described in Supporting Information.
X-ray Structure of [Cu O(CF COO)
Similarly as it was done for the toluene solvate, the geometrical
and/or thermal restraints were applied to the CF groups and
4
3
6
(quin)
4 6 6
]·0.8C H (4).
3
The orientation of the copper basal planes is similar in both
tetramers. Every basal plane is almost perpendicular to two
others and tilted by 25.5° (Cu2/Cu3) or 28.5° (Cu1/Cu4) to the
third plane. This effect is observed for basal planes with copper
one of the quinoline molecules as well as to the partially
occupied benzene sites. Hydrogen atoms were added according
to geometry and not refined. They were omitted for the benzene
1
0282 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
Table 2. Selected Interatomic Distances [Å] and Angles [deg] for 3 and 4
3
4
distance [Å]
angle [deg]
distance [Å]
angle [deg]
Cu(1)-O(61)
1.967(11)
1.969(9)
1.985(12)
2.009(17)
2.339(11)
1.949(8)
1.988(13)
1.992(11)
2.035(14)
2.337(13)
1.929(12)
1.945(12)
1.968(9)
2.014(15)
2.372(11)
1.960(9)
O(61)-Cu(1)-O(5)
90.7(4)
168.7(4)
100.6(4)
84.6(6)
169.9(7)
84.3(6)
92.3(5)
92.1(4)
87.4(5)
96.9(7)
97.9(5)
93.0(4)
168.4(5)
175.0(6)
81.6(5)
87.1(5)
92.8(4)
90.8(5)
92.4(5)
92.1(6)
170.9(5)
97.9(4)
90.7(4)
87.4(6)
83.7(6)
172.1(5)
85.6(4)
96.7(5)
95.0(4)
91.2(6)
90.2(4)
100.2(4)
169.5(5)
174.5(8)
85.1(8)
84.4(8)
150.5(8)
88.2(8)
82.6(8)
32.4(10)
92.9(4)
93.1(5)
86.7(5)
84.5(10)
116.6(8)
107.0(4)
106.6(4)
118.5(5)
118.7(5)
104.2(4)
102.5(4)
Cu(1)-O(5)
1.966(6)
1.968(8)
1.992(7)
2.024(8)
2.297(7)
1.949(9)
1.956(6)
1.984(8)
2.034(11)
2.389(11)
1.949(7)
1.982(8)
1.980(9)
2.022(11)
2.314(11)
2.01(5)
1.966(7)
1.986(9)
1.974(12)
2.13(3)
2.349(12)
3.0199(19)
3.0882(18)
3.3817(21)
3.3960(20)
3.1047(23)
3.1789(28)
O(5)-Cu(1)-O(70)
O(5)-Cu(1)-O(60)
O(70)-Cu(1)-O(60)
O(5)-Cu(1)-N(10)
O(70)-Cu(1)-N(10)
O(60)-Cu(1)-N(10)
O(5)-Cu(1)-O(81)
O(70)-Cu(1)-O(81)
O(60)-Cu(1)-O(81)
N(10)-Cu(1)-O(81)
O(101)-Cu(2)-O(5)
O(101)-Cu(2)-O(90)
O(5)-Cu(2)-O(90)
O(101)-Cu(2)-N(20)
O(5)-Cu(2)-N(20)
O(90)-Cu(2)-N(20)
O(101)-Cu(2)-O(71)
O(5)-Cu(2)-O(71)
O(90)-Cu(2)-O(71)
N(20)-Cu(2)-O(71)
O(5)-Cu(3)-O(80)
O(5)-Cu(3)-O(100)
O(80)-Cu(3)-O(100)
O(5)-Cu(3)-N(30)
O(80)-Cu(3)-N(30)
O(100)-Cu(3)-N(30)
O(5)-Cu(3)-O(110)
O(80)-Cu(3)-O(110)
O(100)-Cu(3)-O(110)
N(30)-Cu(3)-O(110)
N(50)-Cu(4)-O(5)
N(50)-Cu(4)-O(61)
O(5)-Cu(4)-O(61)
N(50)-Cu(4)-O(111)
O(5)-Cu(4)-O(111)
O(61)-Cu(4)-O(111)
N(50)-Cu(4)-N(40)
O(5)-Cu(4)-N(40)
O(61)-Cu(4)-N(40)
O(111)-Cu(4)-N(40)
N(50)-Cu(4)-O(91)
O(5)-Cu(4)-O(91)
O(61)-Cu(4)-O(91)
O(111)-Cu(4)-O(91)
N(40)-Cu(4)-O(91)
Cu(3)-O(5)-Cu(4)
Cu(3)-O(5)-Cu(2)
Cu(4)-O(5)-Cu(2)
Cu(3)-O(5)-Cu(1)
Cu(4)-O(5)-Cu(1)
Cu(2)-O(5)-Cu(1)
88.7(3)
100.4(3)
170.6(3)
161.9(3)
88.5(3)
83.5(3)
95.3(3)
92.9(3)
84.1(3)
102.7(3)
99.0(3)
169.4(3)
91.7(3)
86.2(4)
171.7(5)
83.2(4)
88.1(5)
93.5(3)
91.4(4)
93.3(6)
92.3(3)
98.0(3)
169.7(4)
175.90(4)
87.3(4)
82.3(4)
91.9(3)
90.1(4)
90.6(4)
92.2(5)
156.1(11)
93.3(10)
97.4(3)
75.2(11)
92.9(5)
168.5(5)
29.9(10)
173.7(8)
78.0(8)
92.1(9)
109.1(14)
92.0(3)
91.8(4)
93.0(5)
83.9(8)
108.6(3)
120.8(3)
104.7(3)
104.1(3)
118.7(3)
100.7(3)
Cu(1)-O(5)
Cu(1)-O(101)
Cu(1)-N(30)
Cu(1)-O(111)
Cu(2)-O(5)
Cu(2)-O(100)
Cu(2)-O(70)
Cu(2)-N(20)
Cu(2)-O(91)
Cu(3)-O(81)
Cu(3)-O(90)
Cu(3)-O(5)
Cu(3)-N(10)
Cu(3)-O(60)
Cu(4)-O(5)
Cu(4)-O(110)
Cu(4)-O(80)
Cu(4)-N(40)
Cu(4)-N(50)
Cu(4)-O(71)
Cu(1)-Cu(2)
Cu(1)-Cu(3)
Cu(1)-Cu(4)
Cu(2)-Cu(3)
Cu(2)-Cu(4)
Cu(3)-Cu(4)
O(61)-Cu(1)-O(101)
O(5)-Cu(1)-O(101)
O(61)-Cu(1)-N(30)
O(5)-Cu(1)-N(30)
O(101)-Cu(1)-N(30)
O(61)-Cu(1)-O(111)
O(5)-Cu(1)-O(111)
O(101)-Cu(1)-O(111)
N(30)-Cu(1)-O(111)
O(5)-Cu(2)-O(100)
O(5)-Cu(2)-O(70)
O(100)-Cu(2)-O(70)
O(5)-Cu(2)-N(20)
O(100)-Cu(2)-N(20)
O(70)-Cu(2)-N(20)
O(5)-Cu(2)-O(91)
O(100)-Cu(2)-O(91)
O(70)-Cu(2)-O(91)
N(20)-Cu(2)-O(91)
O(81)-Cu(3)-O(90)
O(81)-Cu(3)-O(5)
O(90)-Cu(3)-O(5)
O(81)-Cu(3)-N(10)
O(90)-Cu(3)-N(10)
O(5)-Cu(3)-N(10)
O(81)-Cu(3)-O(60)
O(90)-Cu(3)-O(60)
O(5)-Cu(3)-O(60)
N(10)-Cu(3)-O(60)
O(5)-Cu(4)-O(110)
O(5)-Cu(4)-O(80)
O(110)-Cu(4)-O(80)
O(5)-Cu(4)-N(40)
O(110)-Cu(4)-N(40)
O(80)-Cu(4)-N(40)
O(5)-Cu(4)-N(50)
O(110)-Cu(4)-N(50)
O(80)-Cu(4)-N(50)
N(40)-Cu(4)-N(50)
O(5)-Cu(4)-O(71)
O(110)-Cu(4)-O(71)
O(80)-Cu(4)-O(71)
N(40)-Cu(4)-O(71)
N(50)-Cu(4)-O(71)
Cu(2)-O(5)-Cu(4)
Cu(2)-O(5)-Cu(3)
Cu(4)-O(5)-Cu(3)
Cu(2)-O(5)-Cu(1)
Cu(4)-O(5)-Cu(1)
Cu(3)-O(5)-Cu(1)
Cu(1)-O(70)
Cu(1)-O(60)
Cu(1)-N(10)
Cu(1)-O(81)
Cu(2)-O(101)
Cu(2)-O(5)
Cu(2)-O(90)
Cu(2)-N(20)
Cu(2)-O(71)
Cu(3)-O(5)
Cu(3)-O(80)
Cu(3)-O(100)
Cu(3)-N(30)
Cu(3)-O(110)
Cu(4)-N(50)
Cu(4)-O(5)
Cu(4)-O(61)
Cu(4)-O(111)
Cu(4)-N(40)
Cu(4)-O(91)
Cu(1)-Cu(2)
Cu(1)-Cu(3)
Cu(1)-Cu(4)
Cu(2)-Cu(3)
Cu(2)-Cu(4)
Cu(3)-Cu(4)
1.970(11)
1.972(12)
2.03(3)
2.15(3)
2.257(12)
3.3702(27)
3.0694(28)
3.1007(29)
3.1413(27)
3.1426(27)
3.3764(27)
atoms joined by carboxylate anion coordinated in equatorial
positions to both centers. In 4 all Cu atoms are shifted by
range from 1.207(17) to 1.264(24) Å, and these bond lengths
exhibit no correlation with the coordination mode and contacts
to copper.
0
.009-0.167 Å from the basal plane toward the axial oxygen
atom. Contrary to 3, such shift is observed also for Cu4 atom
that is bound to disordered quinoline molecule, and it adopts
the smallest (N40) and the biggest values (N50). The angles
around Cu4 atom reveal that its coordination sphere in 4 is less
distorted than in 3. This conclusion is also confirmed by smaller
rmsd (0.031-0.173 Å) of atoms from the basal plane than
observed for 3. Surprisingly, it reaches the largest magnitude
for the basal plane involving the N10 atom and coordination
sphere around Cu1, whereas for the disordered quinoline it is
smaller: 0.076 and 0.160 Å for the N40 and N50 basal planes,
respectively.
The distances between the metal centers forming a tetrahedron
around the central O5 atom vary from 3.020(2) to 3.396(2) Å.
The intermetallic distances as well as the torsion angles
M-O-C-O/O-C-O-M (6.27-46.59°) indicate that those
copper atoms are bound in a syn/syn arrangement. However,
the torsion angles as well as the intermetallic distances are
somewhat larger than found for 3.
Important hydrogen bonds are listed and the crystal packing
is described in Supporting Information.
Comparison of 3 and 4 to Selected Known Trifluoro-
acetato-Bridged Copper Dimers and Tetramers. In dimer 2, the
quinoline nitrogen is in an axial position of Cu, whereas it is
an equatorial ligand in the tetramers 3 and 4. There are no other
All carboxylates are involved in bridging interactions in the
same manner as observed in 3. The distances C-O(carboxylate)
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10283

A R T I C L E S
Ozarowski et al.
significant differences between 3, 4, and 2 as far as the bond
lengths in the copper coordination sphere are concerned. In the
dimer, all oxygen atoms are found in the basal plane and their
6
contacts to copper (1.956(7)-1.990(6) Å) are similar to those
in the square pyramid base in the tetramers. The Cu-N bond
to the axial quinoline is slightly longer (2.109(6) Å) than
equatorial bonds, which results from the Jahn-Teller effect.
The most significant geometric difference is the metal-metal
distance, which equals 2.886(4) Å in 2 and is one of the longest
separations found in copper paddlewheel dimers. However, it
is still shorter by 0.2-0.5 Å than in our tetramer structures.
The Cu-basal plane distance (0.321 Å) is also much larger in
the dimer than that found in the tetramers. The average O-C-O
and Cu-O-C angles in 2, of 129.3(9)° and 124.7(6)°,
respectively, fall within the range observed in 3 and 4 (Table
2
).
A tetramer with formula [Cu
known that differs from our formulas by two hydroxyl groups
2
3 3 2 2
(OH)(CF COO) (quin) ] is
2
-
21
replacing a single O anion. However, it is completely
different structurally from our compounds. The four copper
atoms in it are not chemically identical. There are two inner
metal sites (Cu1 and Cu1′) and two outer ones (Cu2 and Cu2′),
differing in composition of the coordination sphere. The inner
atom polyhedra share a basal edge, whereas the outer atoms
share only one basal corner with the inner copper atoms. That
corner is occupied by an OH group, which creates a H-bond to
a free oxygen atom of the trifluoroacetate ligand that is
monodentately coordinated to Cu2. No monodentate carboxy-
lates are present in 3 and 4.
Single-Crystal Structure Determination of {[Cu
3 4 2
CF COO) (quin)] (quinH )} (2a, Figure 2). The instrumental
2
(OH)-
-
+
(
details are described in the Experimental Section. Crystal data
and structure refinement details are given in Table 3. The crystal
dimensions were 0.240 mm × 0.084 mm × 0.126 mm. The
structure of 2a was solved by direct methods and refined with
Figure 2. (Top) Structure of the tetranuclear anion [Cu
2
OH(CF
3
-
2
-
COO)
the noncoordinated quinH cation forming hydrogen bonds N60-H60· · ·O21
1.95 Å) and N60-H60· · ·F24 (2.82 Å) to the anion.
4 2
(quin)]
in 2a. (Bottom) One-half of the tetranuclear anion and
+
2
a full-matrix least-squares procedure on F using the SHELX-
(
2
0
9
7 program package. Contrary to tetramers 3 and 4, this
structure does not exhibit any significant disorder. Alternative
positions were allowed for one CF group only, which led to a
apical bond by 0.20-0.25 Å. The coordination spheres of both
copper atoms differ significantly. Cu1 is bound only to two
hydroxyl oxygen atoms and to three carboxylate anions, whereas
3
reasonable temperature displacement and an overall model
improvement. Hydrogen atoms were added according to ge-
ometry and refined as riding atoms with fixed temperature
displacement. H atoms of the functional groups (H5 and H60)
were treated exceptionally; their positions as well as the
temperature displacement parameters were refined without any
restraints. They are stable, indicating that the uncoordinated
quinoline molecule and oxygen O5 are protonated.
-
in the Cu2 coordination sphere one OH group is substituted
by a quinoline molecule. In both cases the apical position is
occupied by a carboxylate oxygen atom. In the base of Cu1,
-
both OH groups are mutually in cis positions. Accordingly,
the two carboxylates are also bound in cis positions. Contrary
to this, the quinoline molecule and hydroxyl group are in trans
positions around Cu2 atom and the two carboxylates are also
in trans positions. However, in both cases in the basal plane
there is one bridging and one monodentately bound anion,
whereas the apical position is occupied by another bridging
trifluoroacetate anion. Consequently, the bridging carboxylates
coordinate to an axial position of one copper atom and an
equatorial of another. The basal planes of Cu1 and Cu1#1 are
related by inversion center, while the Cu2 basal plane is almost
perpendicular to them (89.27(9)°). In the basal planes of both
copper coordination polyhedra, the carboxylate oxygen atoms
exhibit the shortest Cu-O separation whereas the hydroxyl
group and quinoline nitrogen atom are bound at a slightly longer
(by 0.03 Å) distance. All four copper atoms lie exactly in one
plane and form a slightly distorted diamond whose short axis
joins Cu1 and Cu1#1 (3.005 Å) while the long one joins Cu2
It is noteworthy that our summary formula and that proposed
6
by Moreland, Cu(CF
3
COO)
2
(quin)(H
O per a tetranuclear unit.
Structure Description. Selected bond lengths and angles are
2
O) (which was assumed
to be monomeric), differ only by 2H
2
-
+
2 3 4 2
listed in Table 4. The {[Cu (OH)(CF COO) (quin)] (quinH )}
complex crystallizes in the triclinic P-1 space group. The
asymmetric unit contains two counterions: one-half of the
tetramer is an anion and one protonated quinoline molecule is
a cation. The second half of the tetramer is generated by the
(
1 - x,-y,1 - z) symmetry operator. The coordination sphere
of both crystalographically independent copper atoms (inner Cu1
and outer Cu2) adopts distorted square pyramid geometry owing
to the Jahn-Teller effect, with a significant elongation of the
and Cu2#1 (6.022 Å). The diamond edges are approximately
(
21) Little, R. G.; Moreland, J. A.; Yawney, D. B. W.; Doedens, R. J.
-
J. Am. Chem. Soc. 1974, 96, 3834–3842.
3.36 Å long. The O5 atoms of OH anions act as µ
3
bridges
1
0284 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
Table 3. Crystal Data and Structure Refinement for
O5 atoms are positioned about 0.6 Å above and below the plane
defined by the corresponding three copper atoms. The fourth
position is occupied by a hydrogen atom with the Cu-O5-H5
angles close to being tetrahedral. Both the temperature displace-
ment parameter of that H atom being twice as high as that
observed for O5 and the possibility of the intramolecular H bond
formation to the noncoordinated O40 atom validate the assump-
tion of the hydroxyl group presence.
{
[Cu
2 3 4 2
(OH)(CF COO) (quin)](quinH)} (2a)
formula
26 2 12 2 9
C H16Cu F N O
formula weight
855.49
293(2)
0.71073
triclinic
temperature [K]
wavelength [Å]
crystal system
space group
P-1 (No. 2)
unit cell dimensions [Å] and [deg]
a [Å]
b [Å]
c [Å]
10.480(2)
13.107(3)
13.399(3)
114.54(3)
92.11(3)
102.31(3)
1619.4(6)
2
There are two distinct groups of carboxylate anions. Two
carboxylates are bidentate, and two others coordinate mono-
dentately. However, the analysis of their environment reveals
lack of additional electron density picks, which could be
considered as protons. Instead, these oxygen atoms are stabilized
R (deg)
ꢀ
(deg)
γ (deg) 3
volume [Å ]
Z
3
by hydrogen bonds formed to a µ hydroxyl group or to the
3
3
protonated quinoline moiety. These interactions are not as strong
as the coordination, and as a result, the respective C-O bonds
in carboxylate groups are short. In 2a, the carboxylate anions
connect each copper atom with two others but never with their
images generated by (1 - x,-y,1 - z) symmetry operator. Thus,
they mediate interactions neither between two Cu1 nor between
two Cu2 atoms. Cu1 and Cu1#1 are joined by two hydroxyl
groups. However, there is no direct bridge between two outer
Cu2 atoms located on the long axis of the diamond. Each of
the copper atoms is coordinated monodentately by one car-
boxylate anion, O21 (Cu1) and O40 (Cu2), which subsequently
D
calc [Mg/m ] [g/cm ]
1.754
1.435
848
absorption coefficient [mm-1]
F(000)
crystal size [mm]
θ range for data collection [deg]
limiting indices/index ranges
0.240 × 0.084 × 0.126
2.43 to 23.26
-8 e h e 11, -14 e k e 14,
-
14 e l e 14
reflections collected/reflections unique 9318/4652 [Rint ) 0.0350]
transmission maximum/minimum
completeness to θ ) 23.26 [%]
refinement method
0.78650/0.92785
99.7
full-matrix least-squares on F
4652/0/503
2
data/restraints/parameters
goodness-of-fit on F2
0.980
a
+
final R indices [I > 2σ(I)]
R1 ) 0.0465, wR2 ) 0.1104
R1 ) 0.0723, wR2 ) 0.1233
0.704 and -0.413
create an intramolecular H bond to N60 of a quinH cation
a
R indices (all data)
and to O5 of the hydroxyl group, respectively. All copper atoms
are coordinated in syn or syn/syn conformation, and the
coordination produces some stress that results in Cu-O-C-O/
O-C-O-Cu torsional angles differing slightly from 0°.
largest diff peak and hole [e A-3]
a
2
2)2/Σ(w|F | ) ]1/2.
2 2
R1 ) Σ|F
o
| - |F
c
|/Σ|F
o
|. wR2 ) [Σw(F
o
- F
c
o
a
Table 4. Selected Bond Lengths [Å] and Angles [deg] for 2a
distance [Å] angle [deg]
Cu(1)-O(20) O(20)-Cu(1)-O(10)
Comparison of 2a with 3, 4, and Little’s Tetramer. The
structure of 2a is substantially different from our tetramers 3
2
1
and 4 and is topologically similar to Little’s tetramer,
[Cu (OH)(CF COO) (quin) . The bond lengths between copper
atoms and µ or µ oxygen atoms are similar, and only two
1.945(3)
1.951(3)
1.962(3)
1.994(3)
2.241(3)
1.940(4)
1.969(4)
1.968(3)
1.995(4)
88.99(15)
170.54(14)
97.54(14)
91.80(14)
175.32(14)
81.14(14)
90.07(15)
89.04(14)
96.84(14)
95.56(13)
131.67(11)
138.33(11)
40.97(9)
Cu(1)-O(10)
Cu(1)-O(5)#1
Cu(1)-O(5)
Cu(1)-O(31)
Cu(2)-O(41)
Cu(2)-O(30)
Cu(2)-O(5)
Cu(2)-N(50)
O(20)-Cu(1)-O(5)#1
O(10)-Cu(1)-O(5)#1
O(20)-Cu(1)-O(5)
O(10)-Cu(1)-O(5)
O(5)#1-Cu(1)-O(5)
O(20)-Cu(1)-O(31)
O(10)-Cu(1)-O(31)
O(5)#1-Cu(1)-O(31)
O(5)-Cu(1)-O(31)
O(20)-Cu(1)-Cu(1)#1
O(10)-Cu(1)-Cu(1)#1
2
3
3
2 2
]
3
4
Cu-OH bonds in 2a are longer by 0.03 Å than other Cu-O or
Cu-OH distances. We observe that the average Cu-N bond
lengths are longer by 0.02-0.03 Å in the µ
4
-oxo tetramers 3
and 4 than in planar copper tetramers with the shortest distance
2
1
found in [Cu
2
(OH)(CF
3
COO)
3
(quin)
2
]
2
This effect is even
Cu(2)-O(11)#1 2.201(4)
clearer in the case of axial Cu-O bonds, which are elongated
by 0.1-0.15 Å in our tetrahedral tetramers. The differences
observed for equatorial Cu-O bonds are not conclusive. We
can state only that in 2a they are 0.1-0.2 Å shorter than in
three other tetramers discussed here. Thus, the main differences
concern the ligand arrangement around the copper atoms.
Tetramer 2a is the only one among the compounds discussed
here in which quinoline molecules are coordinated only to two
outer copper atoms (Cu2 and symmetry-related Cu2#1).
O(5)-Cu(1)#1
O(5)-H(5)
Cu(1)-Cu(2)
1.962(3)
0.90(6)
3.3617(17) O(5)#1-Cu(1)-Cu(1)#1
Cu(1)-Cu(1)#1 3.0047(13) O(5)-Cu(1)-Cu(1)#1
Cu(1)-Cu(2)#1 3.3679(15) O(31)-Cu(1)-Cu(1)#1
Cu(2)-Cu(2)#1 6.0216(31) O(41)-Cu(2)-O(30)
O(41)-Cu(2)-O(5)
40.17(9)
98.17(9)
169.94(17)
94.40(16)
92.71(15)
85.94(18)
85.75(17)
169.16(16)
95.26(19)
91.21(16)
94.88(14)
95.88(17)
117.97(16)
98.86(14)
116.12(15)
O(30)-Cu(2)-O(5)
O(41)-Cu(2)-N(50)
O(30)-Cu(2)-N(50)
O(5)-Cu(2)-N(50)
Magnetic Susceptibility and EPR Spectra of the Tet-
rameric Complexes 3 and 4. The Heisenberg Hamiltonian that
will be used in this paper for interpretation of the magnetic data
for the tetramers is
O(41)-Cu(2)-O(11)#1
O(30)-Cu(2)-O(11)#1
O(5)-Cu(2)-O(11)#1
N(50)-Cu(2)-O(11)#1
Cu(1)#1-O(5)-Cu(2)
Cu(1)#1-O(5)-Cu(1)
H ) J S S + J S S + J (S S + S S + S S + S S )
Cu(2)-O(5)-Cu(1)
1
1
2
2
3
4
3
1
3
1
4
2
3
2 4
Cu(1)#1-O(5)-H(5)
112(3)
101(4)
111(4)
(1)
Cu(2)-O(5)-H(5)
Cu(1)-O(5)-H(5)
This Hamiltonian gives rise in a copper tetramer to one quintet
a
(
S ) 2) level, three triplets (S ) 1), and two singlet states (S )
Atoms generated by symmetry operator -x + 1,-y,-z + 1 are
2
2,23
marked with #1.
0). The spin state energies can be easily evaluated:
forming two equal short bonds to two copper atoms, while the
Cu-O bond to a third copper atom is longer by 0.03 Å. These
(22) Sinn, E. Coord. Chem. ReV. 1970, 5, 313–347.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10285

A R T I C L E S
Ozarowski et al.
ES)2 ) J /4 + J /4 + J
3
1
2
ES)1 ) J /4 - 3J /4, E′
) -3J /4 + J /4,
1 2
1
2
S)1
E′′S)1 ) J /4 + J /4 - J
3
1
2
ES)0 ) J /4 + J /4 - 2J , E′
) -3J /4 - 3J /4
1
2
3
S)0
1 2
(
2)
In the µ
copper atoms from different tetranuclear units, for example, in
4
-oxo tetramers there are no close contacts between
3
the shortest such distance is Cu3-Cu1(0.5 + x,y,0.5 - z)
8
.651(1) Å and Cu4-Cu2(0.5 - x,0.5 + y,z) 8.574(2) Å. The
intertetramer interactions are thus not expected to affect the
magnetic properties. The four copper ions in tetrameric mol-
ecules are joined by the central oxygen bridge and by six
carboxylato bridges. The bridges Cu1-Cu2 and Cu3-Cu4 are
symmetric (see the structure description and Figure 5), while
other four carboxylato bridges are asymmetric. Also, the
Cu1-O5-Cu2 and Cu3-O5-Cu4 angles of 119° are much
Figure 3. Magnetic moment (related to four Cu ions) for the benzene solvate
4. Blue circles are experimental values measured at 5000 G; red dots are
-1
-1
calculated with gaverage ) 2.09, J ) -102 cm , J
1
2
) -39 cm , DS)2 )
-1
-1
-
0.89 cm , ES)2 ) -0.011 cm . Inclusion of the zfs terms in addition to
the Zeeman term is necessary for correct magnetic moment evaluation at
temperatures below 10 K. The inset shows spin state energies. Red, blue, and
black lines represent the quintet, triplets, and singlets, respectively.
larger than other four Cu
07°). Thus, J ) J was set in the Hamiltonian (eq 1).
Strong increase of the effective magnetic moment (related
i
-O5-Cu angles (102°, 104°, 107°,
j
1
1
2
to four Cu(II) ions, Figure 3) with decreasing temperature is an
indication of ferromagnetic coupling in the system. Considering
the magnetic moment of ca. 5.1 µ at the maximum, the
B
population of the S ) 2 ground state must be complete, while
the magnetic moment decrease at the lowest temperatures is
caused by the combined zero-field and Zeeman splitting of the
S ) 2 state becoming comparable to kT.
To reproduce the magnetic moment decrease at the lowest
temperatures, it is necessary to take the Hamiltonian of the
system as a sum of the isotropic part (eq 1) and the part
expressing the anisotropic metal-metal interactions and the
Zeeman interaction in the tetramer:
3
4
4
Haniso
)
S {D }S + µ
B{g }Sk
(3)
∑ ∑
i
ij
j
B
∑
k
i)1 j)i+1
k)1
Figure 4. Magnetization of 4, in multiples of Nµ , measured at 1.8 K
B
(
black circles). The theoretical saturation value equals gS. Blue line:
where {Dij} is a zero-field splitting tensor representing the sum
of the anisotropic exchange and dipole-dipole interactions
magnetization calculated as ꢁB/Nµ with inclusion of the zero-field splitting
using ꢁ from eq 4. Red line: calculated from the Brillouin function including
only the Zeeman term: M ) -gΣm)-2m exp(-gµ
kT). g ) 2.09 was used in both calculations.
B
2
2
B
B
Bm/kT)/Σm)-2 exp(-gµ Bm/
between Cu
i
and Cu
j k k
. {g } is the g-tensor of Cu . The
orientations of {g
k
} (particularly of their z components) are
predictable from the geometric structure, but the orientations
of the {Dij} tensors are not. For the powder magnetic suscep-
tibility calculation it is sufficient to assume isotropic {g} on
each copper atom and to take into account the zero-field splitting
1
Gauss below the working magnetic field of the magnetometer,
B ) 5000 G. Inclusion of the D and E parameters for S ) 2
spin state in addition to the Zeeman term dramatically improved
the magnetic moment fitting for 3 and 4 at temperatures below
of only the ground S ) 2 state. In this way, J
1
2 3
) J and J were
1
0 K (Figure 3), and that effect is even more visible in the
fitted while the D and E parameters for S ) 2 were fixed as
found from EPR (see the EPR section). The average g value
was allowed to vary. The 16 by 16 Hamiltonian matrix was
magnetization measured at 1.8 K (Figure 4). The fitting resulted
in J
-1
-1
1
) J
2
) -97(3) cm , J
3
) -37(1) cm for 3 and J )
1
-1
-1
J
2
) -102(2) cm , J ) -39(1) cm for 4.
To avoid the complexity of the above treatment, one can limit
3
i
diagonalized to find energies E and the molar magnetic
susceptibility was expressed as
the data set to fit only susceptibilities above ∼20 K. The
isotropic Hamiltonian (eq 1) is then acceptable, and susceptibil-
ity can be calculated from the more commonly known formula
1
6
∂
Ei
B
exp(-E /kT)
∑
i
∂
N
B
i)1
ꢁ
) -
(4)
6
1
6
(
2S + 1)(S + 1)S exp(-E /kT)
i i i i
2
2
∑
exp(-E /kT)
∑
i
Ng µ
B i)1
i)1
ꢁ )
(5)
6
3
kT
The derivatives δE
i
/δB were evaluated numerically by
∑
(2S + 1) exp(-E /kT)
i i
i)1
diagonalizing H + Haniso at magnetic fields 1 Gauss above and
where the summations run over two singlet, three triplet, and
one quintet state whose energies are given by eqs 2. When doing
so, essentially the same exchange integrals as above are found.
(
23) Buvaylo, E. A.; Kokozay, V. N.; Vassilyeva, O.Yu.; Skelton, B. W.;
Jezierska, J.; Brunel, L. C.; Ozarowski, A. Inorg. Chem. 2005, 44,
2
06–216.
1
0286 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
Figure 6. Copper coordination and the system of bridges in [Cu
4 4
(µ -
O)(bahped) ](ClO . Black lines indicate the oxygen atoms bridging an
2
4 2
)
equatorial position of one copper atom with an axial position of another
copper atom in the 1-3 and 2-4 copper pairs. The diagram was generated
from cif data in ref 16.
Figure 5. Copper coordination and the system of bridges in 3 and 4. Purple
)
Cu, blue ) N, red ) O, black ) C. Two carboxylate bridges (marked in
green) join an equatorial position of one copper with an equatorial position
of another copper atom, while the four remaining carboxylates join an
equatorial position of one copper atom with an axial position of another.
a b
uses the exchange Hamiltonian H ) -2JS S , was converted
to the notation used in this paper by multiplying it by -2. Our
calculations resulted in JCu1-Cu2 ) -108 (119°), JCu3-Cu4 ) -122
(
)
Cu
119°), JCu1-Cu3 ) -56 (102°), JCu1-Cu4 ) -82 (104°), JCu2-Cu3
1
-55 (107°), and JCu2-Cu4 ) -33 cm (107°). Angles
Both the large ferromagnetic values of the exchange integrals
and the large difference between them appear surprising, as other
i
j
-O5-Cu are given in parentheses. Two interactions,
1
6,18,24-29
JCu1-Cu2 and JCu3-Cu4, involve the symmetric carboxylato bridges,
and their average of -115 cm can be considered to represent
J
µ
4
-oxo copper tetramers are antiferromagnetic.
The
-
1
fitted gaverage value of 2.09 was somewhat smaller than found
from EPR (2.14). Such discrepancies are common.
1
) J in the Hamiltonian (eq 1), while the average value of
2
-
1
-
56 cm for the other four, which are transmitted through the
Estimation of the Exchange Integrals by DFT Calcula-
tions. “Broken symmetry” calculations were attempted by using
asymmetrically bound carboxylate bridges, represents J . The
3
3
0
calculations are in surprisingly good agreement with the results
of magnetic susceptibility fitting; the exchange interactions are
indeed expected to be strongly ferromagnetic and the predicted
the free software package ORCA developed by F. Neese. For
that purpose the tetrameric molecule was simplified by placing
pyridine molecules on the quinoline positions. That transforma-
tion was accomplished by inserting hydrogen atoms on the
vectors joining appropriate carbon atoms in the quinoline rings.
Coordinates from the X-ray structure of 3 were used for all other
atoms, and the structure was not optimized by DFT. Six separate
calculations were performed. In each of them two of the four
copper ions were replaced by zinc, thus leaving a system with
only one exchange interaction between the remaining two copper
3
J1,2/J ratio is large. It is however surprising that the ferromag-
netic interactions are both calculated and found experimentally
with the CuOCu angles as large as 119°. Bera et al. reported
the copper tetramer [Cu
4
(µ
4
-O)(bahped)
2
4 2
](ClO ) , which they
described as containing only a µ
4
-oxo bridge without any
16
bridging ligand along the six tetrahedral edges. However, there
are four bridges in that structure, in which an equatorial oxygen
atom of one copper is 20° away from the axial position of
another copper atom at a distance 2.47 Å (Figure 6). The
magnetic properties were interpreted using the Heisenberg
Hamiltonian with three pairs of equal J values, J , J , and J .
3
1
ions. The calculation utilized the basis Ahlrichs-VDZ and
3
0
polarization functions from the Ahlrichs polarization basis.
In each case the exchange integrals were calculated according
3
2
1
2
3
to the convention J ) 2(EHS - EBS)/(SHS (SHS + 1)), where
HS and BS denote the high-spin state and the broken-symmetry
state, respectively. This convention is equivalent to that of ref
(
Note that eqs 2 are not applicable to this case). J
1
) J
2
) 120
-
1
3
cm was found (converted to our notation), whereas J was
1
6
undetermined but also antiferromagnetic. We performed
broken symmetry calculations for this complex in the same way
as above after replacing two copper atoms by zinc and
33 where exchange coupling in copper paddlewheel dimers was
studied by DFT. The original formula for J from ORCA, which
shortening the CH
34 ) +127 (131°), J13 ) J24 ) -61 (93°), J14 ) J23 ) -49
(108°). The Cu-µ O-Cu angles are in parentheses. The
stronger ferromagnetic interactions are calculated for two copper
pairs, 1-3 and 2-4 for which the µ -oxo angle is the smallest.
The copper atoms in each of these two pairs are joined by two
asymmetric phenoxo-bridges in addition to the µ -oxo bridge.
The µ -oxo bridges 1-2 and 3-4 with the most obtuse angle
transmit strong antiferromagnetic interactions resulting in the
diamagnetic ground state for the complex, in agreement with
the experiment, although only J12 ) J34 are close to those
determined in ref 16.
3
groups in bahped to H. The results are J12
(
(
(
(
(
24) Lines, M. E.; Ginsberg, A. P.; Martin, R. L. Phys. ReV. Lett. 1972,
)
J
2
8, 684–687.
25) Mukherjee, S.; Weyherm u¨ ller, T.; Bothe, E.; Wieghardt, K.; Chaudhuri,
P. Eur. J. Inorg. Chem. 2003, 863–875.
4
26) Atria, A. M.; Vega, A.; Contreras, M.; Valenzuela, J.; Spodine, E.
Inorg. Chem. 1999, 38, 5681–5685.
4
27) Chen, L. Q.; Breeze, S. R.; Rousseau, R. J.; Wang, S. N.; Thompson,
L. K. Inorg. Chem. 1995, 34, 454–465.
4
28) Black, T. D.; Rubins, R. S.; De, D. K.; Dickinson, R. C.; Baker, W. A.
J. Chem. Phys. 1984, 80, 4620–4624.
4
(
(
29) Buluggiu, E. J. Chem. Phys. 1986, 84, 1243–1246.
30) Neese, F. ORCA, an ab initio, Density Functional and Semiempirical
Program Package, Version 2.6-35, 2008; Universit a¨ t Bonn: Bonn,
Germany; free download from http://www.thch.uni-bonn.de/tc/orca/,
registration required.
(
31) (a) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
Some control calculations have also been performed. Thus,
(
b) Ahlrichs et al. Unpublished work. (c) The Ahlrichs auxiliary basis
-1
we found by ORCA J ) 380 cm for 1, while the experimental
sets were obtained from the TurboMole basis set library under
ftp.chemie.uni-karlsruhe.de/pub/jbasen. (d) Eichkorn, K.; Treutler, O.;
Ohm, H.; Haser, M.; Ahlrichs, R. Chem. Phys. Lett. 1995, 240, 283–
-
1
value is 310-340 cm
depending on the method used
(magnetic susceptibility versus EPR intensity measurements, see
2
89. (e) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Theor.
Chem. Acc. 1997, 97, 119–124.
(
32) A. Bencini, A.; Gatteschi, D. J. Am. Chem. Soc. 1986, 108, 5763–
(33) Rodriguez-Fortea, A.; Alemany, P.; Alvarez, S.; Ruiz, E. Chem.sEur.
J. 2001, 7, 627–637.
5
771.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10287

A R T I C L E S
Ozarowski et al.
Figure 7. EPR spectra of 3 (bottom) and of 4 (top) at 15 K and 324 GHz.
The molecular orientations for the resonance lines are marked by X, Y and
Z. The distance between the outer Z lines is 5.48 T in 4 and 5.18 T in 3.
Arrows indicate signals due to frozen oxygen (S ) 1, giso ) 2, D ) 3.57
-1
48
cm ) that is frequently observed below T ) 30 K.
below). For dimer 2 with quinoline shortened to pyridine the
6
-1
-1
calculated J was 350 cm , while 310 cm has been reported.
The calculated J values appear to be imperfect but sensible and
give us more confidence in our results obtained for 3. It should
be mentioned that the exchange coupling in tetranuclear copper
cubanes has been studied by using DFT without resorting to
Figure 8. Resonance field versus frequency dependences observed for 4
at 15 K. Black triangles are the experimental points. The green, blue, and
red lines were calculated for the magnetic field oriented along the molecular
x, y, and z axes, respectively, assuming S ) 2 with the spin Hamiltonian
parameters from Table 5. The group of lines with steepest slope corresponds
3
3,34
the copper substitution by zinc.
EPR Spectra of the µ -oxo Tetramers 3 and 4 (Figure 7).
4
Both 3 and 4 exhibited at low temperatures and high microwave
frequencies complicated and very well resolved EPR spectra
characteristic of S ) 2 spin state. The resolution deteriorated
fast with temperature above 60 K and no spectra of excited
to the “allowed” transitions ∆M
S
) 1, while the slopes of the ∆M ) 2, 3,
S
and 4 transitions are progressively smaller. The black line near the bottom
represents the “minimum magnetic field”, which is due to an off-axial ∆MS
) 4 turning point with an effective magnetic field orientation Θ ) 54°, Φ
23,28,29
triplet states were observed, which is rather typical.
Both
)
45°.
triplet and quintet state spectra in a tetrameric copper complex
were observed in ref 35.
The spin Hamiltonian used to describe the quintet-state spectra
was
simulations improved somewhat when strain was taken into
account. With the magnitudes of J1,2 and J as found from the
3
magnetic susceptibility, no EPR transitions are possible between
the ground S ) 2 state and the excited states. No resonances
within the excited states were identified with the possible
exception of a broad line in the center of the spectra of both 3
and 4 (Figure 7). X-ray structures of both tetramers reveal high
dynamics and disorder, and it may be possible that molecules
freeze into a number of conformations with slightly different
EPR parameters. Also, the zero-field splitting tensor is likely
to be noncoaxial with the g tensor and the Dzialoshinskii-Moriya
interactions may be operative in the system where copper atoms
2
z
2
x
2
y
H ) µ BgS + D{S - S(S + 1)/3} + E(S - S ) +
B
0
0
2
2
4
4
4
B O + B O + B O (6)
4
4
4
4
4
Expressions for the fourth-rank spin operators O i
are given
4
for example in ref 36. The spin Hamiltonian parameters were
found by least-squares fitting of the resonance field versus
microwave frequency dependencies measured over a wide
frequency range (Figure 8 and Figure S2, Supporting Informa-
tion). The fitting required usage of the fourth-order terms B
The fit was quite good, and the resonance fields for both the
allowed” transitions and the “forbidden” transitions with ∆M
2, 3, and 4 were very well simulated.
The simulation of the powder spectra was less satisfactory,
as far as the relative intensities are concerned. One of the reasons
for the problems may be the very peculiar line width pattern.
The linewidths vary with orientation and also are different for
various transitions at an orientation. For example, the outer Z
i
4
.
24
are not related by the inversion center but both effects cannot
be further investigated in the absence of large single crystals
suitable for high-field EPR. Nevertheless, considering the
goodness of the fit shown in Figures 7 and S2 in Supporting
Information, and the fact that the resonance fields for the
“
)
S
“forbidden” lines were well reproduced, we conclude that the
g components as well as the magnitudes of D and E have been
determined with good certainty. The spin Hamiltonian param-
eters for 3 and 4 are given in Table 5. The benzene solvate 4
exhibits larger D than 3, while E is smaller in 4 than in 3. The
g components are equal for both complexes within the error
limits.
lines corresponding to the transitions between the M
T 1 and -2 T -1 are much broader than the inner 1 T 0 and
1 T 0 transitions indicating a “D-strain” in the system. Powder
S
states 2
-
EPR spectra of copper tetramers have been occasionally
2
3,28,29,35
(
(
34) (a) Ruiz, E.; Alvarez, S.; Cano, J.; Polo, V. J. Chem. Phys. 2005,
23, 164110. (b) Tercero, J.; Ruiz, E.; Alvarez, S.; Rodriguez-Fortea,
reported, but successful interpretations are very rare.
1
28
Black et al. investigated a highly symmetric tetranuclear
copper triphenylphosphine oxide complex Cu OCl (TPPO) and
A.; Alemany, P. J. Mater. Chem. 2006, 16, 2729–2735.
35) Prins, R.; Degraaff, R. A. G.; Haasnoot, J. G.; Vader, C.; Reedijk, J.
J. Chem. Soc. Chem. Commun. 1986, 18, 1430–1431. Mialane, P.;
Duboc, C.; Marrot, J.; Rivi e` re, E.; Dolbecq, A.; S e´ cheresse, F.
Chem.sEur. J. 2006, 12, 1950–1959.
4
6
4
interpreted its EPR spectra by using a spin Hamiltonian
containing only the Zeeman interaction and the fourth-order
0
4
cubic zero-field splitting term B
4
(O
4
+5O
4 4
) with B ) 0.0044
(
36) Abragam, A.; Bleaney, B. In Electron Paramagnetic Resonance of
Transition Ions; Dover Publications, Inc.: New York, 1986.
-
1
cm . In our case, large D and E parameters were found in
1
0288 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
a
Table 5. Quintet-State (S ) 2) EPR Parameters for Tetrameric Complexes 3 and 4
-1
-1
0
4
-1
2
4
-1
4
-1
g
x
g
y
g
z
D, cm
E, cm
B
, cm
B
, cm
4
B , cm
3
4
2.168(5)
2.169(3)
2.173(5)
2.175(1)
2.066(1)
2.067(1)
-0.827(2)
-0.875(1)
-0.114(4)
-0.049(1)
0.000(1)
0.000(3)
-0.0114(4)
-0.009(1)
0.0012(4)
-0.006(1)
a
The fitting procedure and estimation of parameter errors are explained in Supporting Information; see Figure S2.
addition to the fourth-order terms. This is not surprising because
the molecular symmetry of 3 and 4 is lower than of the TPPO
complex. In the latter, the energy difference between the lowest
and the highest level of the quintet state at the zero magnetic
g
mR and gni are the g
“m” and “n” in the interacting pair, dγi are elements of the matrix
that transforms g axes into the g axes, and σ are the direction
x y z
, g , and g components for two ions
n
m
R
cosines of the vector r joining the atoms “m” and “n”, expressed
in the g axes of atom “m”. All six tensors were then transformed
into a common system of coordinates and added according to
formula (9) that relates the zero-field splitting tensor components
expressed in terms of the spins of separate ions and those
-1
-1
field is 0.5 cm , while it is about 3.5 cm in 3 and 4. D )
-
1
-
0.353 cm was found in a ferromagnetic tetrameric copper
23
-1
cubane causing an overall zero-field splitting of 1.4 cm . The
D and E parameters in 3 and 4 are exceptionally large and must
be due to the anisotropy of the exchange interactions, since the
dipolar contribution is negligible (see below). It has to be
emphasized here that the EPR spectra of 3 and 4 as well as of
3
7
appropriate for the total spin S ) 2:
1
12
(k)
D (S ) 2) )
Dij
(9)
ij
∑
3
a (below) could not be studied using standard X-band or even
k
Q-band EPR because of the large zero-field splitting.
The g values are very different from those expected for a
where k runs over the six interactions 1-2, 1-3, 1-4, 2-3, 2-4,
and 3-4.
mononuclear Cu(II) with tetragonal symmetry (g
z
≈ 2.2-2.4;
1
,3,8,36
g
x,y ≈ 2.02-2.08),
which is the result of the single ion g
The resulting matrix was diagonalized yielding Dxx
0.0125 cm , Dyy ) 0.0097 cm , Dzz ) 0.0019 cm .
Application of the relations between D, E, and the ZFS tensor
)
tensors not being aligned in the tetranuclear system. There is
no experimental way to determine the g components of
individual ions in an exchange coupled system. It is possible,
however, to use the molecular structure and simple assumptions
about Cu(II) g tensors to explain the results as follows. Using
the X-ray structure of 3, the coordinate systems for each copper
ion were set up with the z axis for each copper ion assumed
perpendicular to the least-squares plane of its equatorial ligands.
The x axis was perpendicular to z and to the vector joining Cu
and the central oxygen and finally y was perpendicular to both
-1
-1
-1
-
components
D ) (2D - D - D )/2, E ) (D - D )/2
zz
xx
yy
xx
yy
(
10)
-1
yielded finally Ddipole(S ) 2) ) -0.0183 cm . This is only
about 1/45 of D as determined from EPR, indicating an
overwhelming contribution of the anisotropic exchange interac-
tions to the zero-field splitting in the tetranuclear molecule.
x and z. The tensors for the four copper atoms, g
assumed axial, i.e., g ) g , and equal to each other, except for
their orientation in space. The g , g , and g tensors were rotated
to the system of coordinates of g , summed according to formula
below and diagonalized to obtain the main values and main
1
-g
4
, were
x
y
-1
Calculated Edipole(S ) 2) is 0.0039 cm , 29 times less than the
2
3
4
experimental E value in 3.
1
Magnetic Properties of the Blue Complex 2a. This blue
substance is different from the blue product 3a forming from
the tetramer and exhibits no discernible EPR spectrum at any
frequency and temperature, which appears to be due to enormous
line width. Magnetic susceptibility of the blue compound was
measured revealing antiferromagnetic exchange interactions
which were much weaker than in the original yellow-green
dimer 2. Before fitting, the monomer susceptibility was sub-
tracted from experimental data. The monomer susceptibility was
calculated from
7
3
7
axes of g in the S ) 2 state.
gS)2 ) (g + g + g + g )/4
(7)
1
2
3
4
When g
z
) 2.31 was assumed implying that g
x
) g
y
) 2.05
(
2
2
the average g must be equal to that found from EPR for S )
) then the resulting three g components for S ) 2 were 2.166,
.180 and 2.065, resembling very closely the values determined
from EPR. gx,y ) 2.05 and g ) 2.31 are reasonable for a copper
z
ion having three oxygen atoms and one nitrogen in its equatorial
plane and may be used as estimates of the single-ion g values
in 3. The g-tensor main axis corresponding to g ) 2.065 lies
ꢁmono ) (Ngµ /2B)(x - 1/x)/(x + 1/x)
(11)
B
where x ) exp(gµ
B
B/2kT) and B is the magnetic induction of
the magnetometer (5000 G).
along the approximate S
4 4
axis of the distorted Cu tetrahedron,
at a 1.2° angle from the O5-C102 vector.
Using the Curie law to calculate the monomer susceptibility
at the lowest temperatures would not be proper because the
Dipolar Contribution to the ZFS Parameters. Formula 3 in
ref 38 was used to calculate dipolar tensors for each of the six
magnetic dipole-dipole interactions in the tetrameric system:
-
1
Zeeman splitting (∼0.47 cm at 5000 G) is comparable to kT
-
1
(1.4 cm at 2 K). The procedure was only partially successful
as the remaining susceptibility did not approach zero at the
lowest temperatures in the manner the antiferromagnetic sus-
ceptibility should (Figure 9). The effect of the monomer
contribution removal on the susceptibility at higher temperatures
was very small, and data below 20 K were subsequently not
fitted. Judging from the molecular structure (Figures 2 and 10),
2
B
3
DRγ ) gmR
{
g d (d - 3σR
∑i ni γi Ri
d σ }µ /r
(8)
∑j ji
j
(
37) Bencini, A.; Gatteschi, D. In EPR of Exchange Coupled Systems;
Springer Verlag: Berlin, Heidelberg, 1990; p 118.
1 2 3
the Hamiltonian (eq 1) with all J , J , and J being different is
(38) Carr, S. G.; Smith, T. D.; Pilbrow, J. R. J. Chem. Soc., Faraday Trans.
2
1974, 70, 497–511.
appropriate for fitting of the magnetic susceptibility.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10289

A R T I C L E S
Ozarowski et al.
Figure 11. Blue: EPR spectrum of 3a. Asterisks indicate resonances due
to a monomeric copper species. Sharp “perpendicular” part of the monomer
signal is cut off to make the triplet spectrum better visible. Red: Spin-
triplet spectrum simulated with g
-1.636 cm , E ) -0.0187 cm .
x
) 2.049, g
y
) 2.263, g ) 2.112, D )
z
-
1
-1
Figure 9. Magnetic susceptibility (per 1 Cu) of the blue tetramer 2a. Blue
circles ) original data points; red circles ) susceptibility after subtraction
of 0.28% of monomeric contamination. Black line was calculated with J
1
-
1
)
J
2
) 0, J
3
) 81 cm , g ) 2.17 (see text). Green circles are the data for
green dimer 2 from ref 6. Solid line was calculated for a dimer with J )
-
1
6
3
10 cm and g ) 2.27.
Figure 12. Magnetic susceptibility of 3a (blue circles). Red circles:
susceptibility after subtraction of 0.93% of monomeric impurity. Solid line
Figure 10. Copper coordination and the bridge system in 2a.
-1
was calculated with J
1
) J
2
) J ) 88 cm and gaverage ) 2.14. The inset
3
shows energies of the spin states. Red, blue, and black lines represent the
quintet, triplets, and singlets, respectively.
J
2
should be small because there is no direct bridge between
Cu3 and Cu4 and was fixed to zero. Basing on results of ORCA
calculations for this compound the J (Cu1-Cu2 in Figure 10)
value was also fixed to zero (see below). The fit then resulted
Blue Product (3a) Forming from the Tetramer (3). Samples
of tetramer 3 kept for several days in humid air changed color
from green to blue and increased in weight by 3%. The blue
compound exhibited EPR spectra characteristic of a triplet state
1
-1
in J ) 81 cm , g ) 2.17 (g could not be determined
3
-
1
with g
-0.019 cm . A weak spectrum of monomeric copper(II)
with g ) g ) 2.048, g ) 2.263 was also observed (Figure
1). No crystals suitable for the structure determination could
be obtained.
x
) 2.049, g
y
) 2.263, g ) 2.112, D ) -1.636 cm , E
z
independently because 2a exhibits no EPR). The model with
only four equal exchange integrals in the ring containing four
copper ions appears to be fully adequate.
-
1
)
x
y
z
1
ORCA calculations were performed as described above, i.e.,
by replacing two copper atoms by zinc and evaluating the
exchange integral between the remaining two copper atoms.
Antiferromagnetic J values were found: the interaction between
two closest copper atoms J12 was the smallest, 3.5 cm ,
followed by four 12 cm- interactions in the ring Cu1-Cu3-
Cu2-Cu4, while J34 ) 25 cm- was the largest. A small J12
magnitude can be rationalized since the Cu1-O-Cu2 angle of
This spectrum broadened very quickly with increasing
temperature and was not seen above 80 K. The triplet spectrum
also disappeared below 10 K, leaving behind only the monomer
signals.
Magnetic susceptibility of the blue compound in Figure 12
shows an antiferromagnet contaminated by a monomer. Before
fitting, the monomer susceptibility was subtracted from experi-
mental data as described above. A 0.93% monomer content can
be estimated from the susceptibility over the range 2-10 K;
this was subtracted from all experimental points and data below
-
1
1
1
9
8.6° lies on the borderline 97-98° separating the ferromagnetic
and antiferromagnetic interactions in dihydroxo-bridged copper
3
9
dimers. As for J34, it is not clear to us whether a “direct”
Cu3-Cu4 exchange should be considered at all when the
exchange pathways involve the inner Cu1 and Cu2 atoms. The
large value of J34 derived from DFT may be associated with
the calculation method, in which the intervening Cu1 and Cu2
atoms were replaced by zinc. In this context it is worth
mentioning that an exchange integral of 35 cm- was found
experimentally in a dimer containing two Cu ions bridged by
2
0 K were subsequently not fitted. The monomer was also seen
in EPR (Figure 11). The susceptibility has a maximum at Tmax
80 K. If the complex were a dimer then the J value would
)
-1
be J ) 1.6 kTmax or 90 cm , but then the susceptibility would
-6
have to be much higher (2600 × 10 cgs emu at the maximum
1
-6
with g ) 2.1) than observed (1700 × 10 ), and the dimer model
has to be rejected, although this compound of unknown structure
exhibits triplet EPR spectra.
4
0
two OZnO groups.
Fitting under assumption that the complex was still tetrameric
(
(
39) Crawford, V. H.; Richardson, H. W.; Wasson, J. R.; Hodgson, D. J.;
Hatfield, W. E. Inorg. Chem. 1976, 15, 2107–2110.
was more successful (Figure 12). The relatively best fit was
-1
achieved with all six J values equal to 88 cm (by setting J
1
40) Buvaylo, E. A.; Kokozay, V. N.; Vassilyeva, O.Yu.; Skelton, B. W.;
Jezierska, J.; Brunel, L. C.; Ozarowski, A. Chem. Commun. (Cam-
bridge) 2005, 39, 4976–4978.
)
J
2
) J
3
in eqs 2) and with fixed gaverage ) 2.14, as found
from EPR. Fitting the magnetic susceptibility of an antiferro-
1
0290 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009

Binuclear and Tetranuclear Copper Trifluoroacetates
A R T I C L E S
magnetic tetramer is not easy and results less satisfactory than
in Figure 12 are commonly seen in the literature. This blue
complex thus appears to be tetranuclear but with drastically
altered magnetic properties compared to its parent green
tetramer. The exchange interactions are now antiferromagnetic
and the ground state is a singlet with S ) 0. The observed triplet
spectrum comes from an excited S ) 1 state lying about 88
-1
cm higher. Crystal structures of 2a, 3, and 4 show high
flexibility of the carboxylates: they may be bridging, chelating,
or dangling while IR spectra indicate that the µ -oxo bridge in
4
the blue complex under discussion is retained. While it is clear
that the bridge structure must have changed, no sufficient
evidence was gathered in this case to propose a structure for
this species.
Figure 13. Powder EPR spectrum of 1 at 80 K and 406.4 GHz. Blue:
experimental; red: simulated with parameters from Table 6. Broad signals
of a monomeric impurity are seen between the S ) 1 state signals. Hyperfine
-
4
-1
structure due to two copper nuclei (A ) 67 × 10 cm , half of that
z
Magnetic and EPR Properties of the Binuclear Com-
expected for a monomeric Cu complex with four equatorial oxygen atoms)
appears on the “Z” resonances.
plexes. [Cu(CF
of 1 has not been reported and was therefore measured here
Figure S5, Supporting Information). The exchange integral
3
COO)
2
·(CH
3 2
CN)] (1). Magnetic susceptibility
5
parameters for 1 and 2 found at 50-80 K from the high-field
EPR spectra (Figures 13 and S4 in Supporting Information) are
shown in Table 6. |D| is substantially larger than in nonhalo-
(
magnitude J, corresponding to the spin Hamiltonian
-
1
genated copper carboxylates (∼0.33 cm , Table 6). Single-
crystal spectra were recorded with ν ) 432 GHz at 80 and 50
H ) JS1S2
(12)
-
4
K. The triplet state population at 50 K is only about 2 × 10 ,
and hyperfine splitting due to two copper nuclei clearly appears
because the paramagnetic molecules are diluted in the diamag-
netic bulk of the sample. Since the spin Hamiltonian parameters
g and D were known from powder spectra, the molecular
orientation Θ ) 29° in the single-crystal spectrum in Figure
was found from least-squares fitting. With gaverage fixed at 2.18
as determined by EPR, J ) 312 cm-1 was obtained. The
temperature-independent paramagnetism (TIP) for copper was
-6
fixed at 60 × 10 cgs units.
A relatively high content of paramagnetic impurity (ca. 3.4%)
was also revealed. This is likely associated with the hygroscop-
icity of this substance. Measurements of the single-crystal EPR
signal intensity was also performed to find J independently. The
high-field “perpendicular” transition centered at 4910 G at the
X-band frequency was used for this purpose.
1
4 could be determined from the resonance positions.
Triplet spin states give rise to two EPR transitions at each
orientation of a molecule versus the magnetic field (see the X,
Y and Z pairs of lines in Figures 13 and S4 and S6 in Supporting
Information). The sign of D can be determined by measuring
the relative intensity of the high-field and the low-field signal
in such a pair at very high microwave frequency. If D is negative
then the low-field line should be stronger than the high-field
The EPR signal intensity was fitted with the equation
I ) (C/T) exp(-J/kT)/(1 + 3 exp(-J/kT))
(13)
8
where C is a proportionality constant. The EPR intensity is
line at Θ < 60°, as is observed. The measured intensity ratio
2
2
8
proportional to the magnetic susceptibility (taking C ) Ng µ
B
/k
of 1.53 at 50 K is very close to the theoretical value of 1.50.
converts eq 13 to the magnetic susceptibility referred to one
Cu atom) but is not affected by paramagnetic contaminants
Similar relations were observed for 2. Negative sign of D in
dimeric perfluorinated copper carboxylates is therefore con-
firmed, and there is no reason to expect it to be positive in any
other copper paddlewheel dimers. The sign of D plays a crucial
role in estimating the anisotropic exchange-related contribution
(which typically give signals close to g ) 2 or 3400 G in X-band
EPR), nor does it require correction for ligand diamagnetism
and temperature independent paramagnetism of copper. How-
ever, the procedure is quite cumbersome owing to a need of
single crystal measurements followed by double integration of
the spectra (see Figures S5, S5a in Supporting Information).
Fortunately, 1 forms very large crystals of high quality
exceptionally easily. J of 339(8) cm- obtained in this way was
in reasonable agreement with the magnetic susceptibility data.
EPR Spectra and the Sign of D in Dimeric Complexes 1
and 2. EPR spectra of binuclear complexes are described by
the spin Hamiltonian
8
to the observed zero-field splitting, and since most literature
reports appear to be incorrect due to the assumption of a positive
sign, we present some EPR parameters and recalculated Dexchange
data in Table 6.
The exchange-related part of D shows correlation with the
electron-withdrawing strength of substituents in the carboxylate
1
chain, and the increase in D
g components, as expected.
exchange
,8,41-45
parallels the increase in the
Dexchange is related (through
1
(
(
(
(
41) Morosin, B.; Hughes, R. C.; Soos, Z. G. Acta Crystallogr. 1975, B31,
2
Z
2
x
2
y
7
62–770.
H ) µ BgS + D{S - S(S + 1)/3} + E(S - S )
B
42) Muto, Y.; Nakashima, M.; Tokii, T.; Suzuki, I.; Ohba, S.; Steward,
O. W.; Kato, M. Bull. Chem. Soc. Jpn. 2002, 75, 511–519.
43) Ross, P. K.; Allendorf, M. D.; Solomon, E. I. J. Am. Chem. Soc. 1989,
111, 4009–4021.
(
14)
Although positive sign of the zero-field splitting parameter
D in binuclear copper carboxylates has been reported in
a recent high-field EPR study of copper acetate
monohydrate dimer and its pyrazinate analogue showed that D
is in fact negative. High-field EPR measurements have thus
been performed in this work on single crystals of 1 and 2 to
determine the sign of D. Powder EPR spectra of antiferromag-
44) Ozarowski, A.; Reinen, D. Inorg. Chem. 1986, 25, 1704–1708.
4
1-43
(45) Gribnau, M. C. M.; Keijzers, C. P. Inorg. Chem. 1987, 26, 3413–
literature,
3
414.
(46) Brown, G. M.; Chidambaram, R. Acta Crystallogr. 1973, B29, 2393–
8
2403.
(
(
47) Nakagawa, H.; Kani, Y.; Tsuchimoto, M.; Ohba, S.; Matsushima, H.;
Tokii, T. Acta Crystallogr. C 2000, 56, 12–16.
48) Pardi, L. A.; Krzystek, J.; Telser, J.; Brunel, L. C. J. Magn. Reson.
2000, 146, 375–378.
8
netic dimers are not suitable for that purpose. The EPR
J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009 10291

A R T I C L E S
Ozarowski et al.
Table 6. Spin Hamiltonian Parameters for Dimeric Complexes
4
-1
4
-1
4
-1
4
-1
complex
g
x
g
y
g
z
10 D, cm
10 E, cm
rCu-Cu, Å
10 Ddipole, cm
10 Dexchange, cm
a
46
[
[
[
[
[
[
Cu(CH
3
COO)
COO)
2
H
2
2
O]
(pyr)]
CN)]
(quin)] (80 K)
2
(80 K)
2.0545(3) 2.0792(2) 2.3637(2) -3350(10)
2.0608(2) 2.0622(2) 2.3493(3) -3280(10)
2.0709(2) 2.0752(2) 2.3905(3) -4096(9)
2.0836(3) 2.0836(3) 2.4115(4) -4330(15)
2.0578(4) 2.0774(4) 2.3618(5) -3188(17)
2.0699(5) 2.0699(5) 2.3697(6) -3332(26)
-103(1)
0(10)
0(10)
0(10)
-83(9)
0(20)
2.616
41
-1710
-1760
-1610
-1440
-1860
-1840
-1640
-1520
-2490
-2890
-1328
-1490
a
,
b
Cu(CH
3
2
(80 K)
(80 K)
2.58
5
Cu(CF
Cu(CF
3
COO)
COO)
2
(CH
3
2
2.766
2.886
2.618
2.632
c
6
3
2
2
d
,
e
47
47
Cu((CH
Cu((CH
3
)
)
3
SiCH
SiCH
2
COO)
COO)
2
]
2
n
(280 K)
e
3
3
2
(py)] (280 K)
2
a
Spin Hamiltonian parameters from ref 8; all others are from this work and were determined from high-field EPR by global fitting of spectra
b
recorded at 4 different frequencies, around 112, 224, 324, and 406 GHz. See also a note below Figure S2 in Supporting Information. pyr: pyrazine; py:
pyridine; quin: quinoline. Significantly incorrect g
X-Band EPR, where |D| is comparable to the microwave quantum energy. The complex forms a chain of paddlewheel entities, but EPR spectra of
isolated S ) 1 state are observed. Complex was prepared as described in ref 47 for comparative purposes, and no data other than EPR were collected.
See Figure S6 in Supporting Information.
c
x
) g ) 2.17 reported in ref 6 illustrates difficulties in extracting spin Hamiltonian parameters from
y
d
e
temperature magnetic susceptibility and magnetization data for
our ferromagnetic tetramers 3 and 4. The dipolar contribution
to the zfs parameters in 3 and 4 was found to be negligible,
whereas it is of comparable magnitude to the exchange-induced
zfs in copper paddlewheel dimers (Table 6 and ref 8). The
copper(II)-trifluoroacetate-quinoline system is unusually rich
with at least five different tetranuclear complexes known at this
6
time (this work and ref 21), plus one binuclear compound. In
addition, at least one synthetic avenue leading to a nonmono-
Figure 14. (Top) Single-crystal spectrum of 1 recorded at 50 K with ν )
32.0 GHz. (Center) Integrated signals. (Bottom) Doubly integrated signals.
meric complex was abandoned in this work. This abundance
of compounds is probably due to the possibility of formation
of extensive hydrogen bond networks that stabilize various
structures. Also, bonds between the carboxylate oxygen atoms
and copper are weaker than in nonhalogenated carboxylate
complexes providing additional flexibility to the system. Finally,
negative sign of the zfs parameter D was found in binuclear
acetonitrile and quinoline adducts of copper trifluoroacetate (1
and 2), and it is postulated that D is negative in all copper
paddlewheel dimers.
4
Signal integration was needed to extract the real intensity ratio of the low-
field line to the high-field line (1.53) because the line widths are considerably
different. The molecular orientation Θ ) 29°, meaning that the magnetic
field lies 29° from the Cu-Cu direction, could be inferred from the
resonance positions.
the spin-orbit coupling) to the exchange interactions in excited
states of a dimeric molecule:
2
2
2
z
2
2
2
y
Dexchange ) [J(x - y , xy)∆ - 2J(x - y , z)∆ -
2
2
2
2
J(x - y , yz)∆ ]/32 (15)
Acknowledgment. This work was supported by the NHMFL,
which is funded by the NSF through Cooperative Agreement No.
DMR-0654118, the State of Florida, and the DOE and by the
Ministry of Science and Higher Education of Poland through grant
N204 049 31/1376. J.J. thanks the Wrocław University for support.
x
2
2
where ∆
z
) g
z
- 2.0023, etc., J(x - y ,n) represent the triplet-
singlet separations in excited dimer states in which one of the
copper atoms is in its electronic ground state dx2
y2, while
Increased magnitudes
-
1,8,41-45
another one is in an excited state n.
of both the g components and Dexchange in fluorocarboxylates
compared to plain carboxylates reflect a reduced covalency of
the Cu-Ocarboxylate bonds in the former ones.
Supporting Information Available: CIF files for 2a, 3, and
. Scheme illustrating transformations of compounds studied
in this work. Energy levels and EPR transitions in the S ) 2
spin state of 3. Resonance fields versus microwave frequency
plot for 3 and note on parameter error estimation. Simulation
of a high-field EPR spectrum of 4. Solid-state conversion 2 T
4
Conclusions
Two new ferromagnetic and two antiferromagnetic (one of
unknown structure) tetranuclear copper complexes were inves-
tigated by high-field EPR and magnetic susceptibility measure-
ments. Broken symmetry DFT calculations were performed for
tetramers of known structure to estimate the exchange integrals.
Calculations correctly predicted the type of magnetism (ferro
versus antiferro) in each of the three cases attempted. The
calculated exchange integral magnitudes were in a semiquan-
titative agreement with the experimental data in both tetramers
and dimers. Knowledge of the zero-field splitting parameters,
which were determined from the high-field, high-frequency EPR,
proved to be crucial for the interpretation of the lowest-
2
a followed by EPR. Exchange integral determination for 1 from
magnetic susceptibility and EPR signal intensity. Experimental
and simulated EPR spectra of [Cu((CH SiCH COO) and
[Cu((CH ) SiCH COO) (py)] . ORTEP plot for 4. Torsional
3
)
3
2
2 n
]
3
3
2
2
2
angles and hydrogen bonds in 2a, 3, and 4 and stacking
interactions in 3 and 4. Shortened output of ORCA calculations.
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA902695Y
1
0292 J. AM. CHEM. SOC. 9 VOL. 131, NO. 29, 2009