Electronic tuning of β-diketiminate ligands with fluorinated substituents: Effects on the O2-reactivity of mononuclear Cu(I) complexes

Article abstract of DOI:10.1039/b609939d

Copper(i) complexes with the β-diketiminate ligands HC{C(R)N(Dipp)}{C(R′)N(Dipp)}^- (Dipp = C₆H ₃ⁱPr_2-2,6; L¹, R = CF₃, R′ = CH₃; L², R = R′ = CF₃) have been isolated and fully characterized. On the basis of X-ray structural comparisons with the previously reported complex LCu(CH₃CN) (L = HC{C(CH₃)N(Dipp)}₂^-), the ligand environments at the copper centers in the analogous nitrile adducts with L¹ and L² impose similar steric demands. L¹Cu(CH₃CN) reacts instantaneously at low temperature with O₂ to form a thermally-unstable intermediate with an isotope-sensitive vibration at 977 cm^-1 (928 cm^-1 with ¹⁸O₂), in accord with the peroxo O-O stretch associated with side-on coordination for LCu(O ₂). However, L²Cu(CH₃CN) is unreactive toward O₂ even at room temperature. Evaluation of the redox potentials of the nitrile adducts and the CO stretching frequencies of the carbon monoxide adducts revealed an incremental adjustment of the electronic environment at the copper center that correlated with the extent of ligand fluorination. Furthermore, theoretical calculations (DFT, CASPT2) predicted that an increasing extent of Cu(ii)-superoxo character and end-on coordination of the O ₂ moiety in the Cu/O₂ product (L² > L ¹ > L) are accompanied by increases in the free energy for the oxygenation reaction, with L² unable to support a Cu/O₂ intermediate. Calculations also predict the 1:1 Cu/O₂ adducts to be unreactive with respect to hydrogen atom abstraction from hydrocarbon substrates on the basis of their stability towards both reduction and protonation. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b609939d

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Electronic tuning of b-diketiminate ligands with fluorinated substituents:
effects on the O2-reactivity of mononuclear Cu(I) complexes†‡
Lyndal M. R. Hill, Benjamin F. Gherman, Nermeen W. Aboelella, Christopher J. Cramer and
William B. Tolman*
Received 12th July 2006, Accepted 14th August 2006
First published as an Advance Article on the web 31st August 2006
DOI: 10.1039/b609939d
−
Copper(I) complexes with the b-diketiminate ligands HC{C(R)N(Dipp)}{C(Rꢀ)N(Dipp)} (Dipp =
C6H3 Pr2–2,6; L1, R = CF3, Rꢀ = CH3; L2, R = Rꢀ = CF3) have been isolated and fully characterized. On
i
the basis of X-ray structural comparisons with the previously reported complex LCu(CH3CN) (L =
HC{C(CH3)N(Dipp)}2−), the ligand environments at the copper centers in the analogous nitrile
adducts with L1 and L2 impose similar steric demands. L1Cu(CH3CN) reacts instantaneously at low
temperature with O2 to form a thermally-unstable intermediate with an isotope-sensitive vibration at
977 cm−1 (928 cm−1 with 18O2), in accord with the peroxo O–O stretch associated with side-on
coordination for LCu(O2). However, L2Cu(CH3CN) is unreactive toward O2 even at room temperature.
Evaluation of the redox potentials of the nitrile adducts and the CO stretching frequencies of the carbon
monoxide adducts revealed an incremental adjustment of the electronic environment at the copper
center that correlated with the extent of ligand fluorination. Furthermore, theoretical calculations
(DFT, CASPT2) predicted that an increasing extent of Cu(II)–superoxo character and end-on
coordination of the O2 moiety in the Cu/O2 product (L2 > L1 > L) are accompanied by increases in the
free energy for the oxygenation reaction, with L2 unable to support a Cu/O2 intermediate. Calculations
also predict the 1 : 1 Cu/O2 adducts to be unreactive with respect to hydrogen atom abstraction from
hydrocarbon substrates on the basis of their stability towards both reduction and protonation.
to render 3 and 4 relatively unreactive,11 limiting efforts to test
the hypothesized involvement of 1 : 1 Cu/O2 species in organic
substrate hydroxylations (e.g. in PHM).12 We therefore asked: How
will decreasing the electron donation of the b-diketiminate ligand
influence the Cu(I)/O2 reactivity?
Introduction
A key step in many oxidation reactions of biological and industrial
importance is the binding and activation of dioxygen at a single
copper center to yield 1 : 1 Cu/O2 species.1,2 Understanding
how ligand structural effects control the geometries, electronic
structures, and reactivity of such species is a primary objective of
current research aimed ultimately at obtaining mechanistic insight
into oxidation catalysis.1,3 Structurally characterized examples of
1 : 1 Cu/O2 species are limited to a trapped form of peptidylglycine
a-hydroxylating monooxygenase (PHM)4 and the synthetic com-
plexes 1–4 (Fig. 1).5–8 X-ray crystallographic, spectroscopic, and
theoretical data support Cu(II)–superoxide formulations for 1 and
2,5,6 whereas 3 and 4 have significant Cu(III)-peroxo character.7–10
The latter electronic structure derives from the nature of the b-
diketiminate and anilido-imine supporting ligands, which through
powerful electron donation stabilize the high oxidation state of
the metal ion and drive the reduction of the dioxygen moiety
to the peroxide level. Unfortunately, these effects also appear
Department of Chemistry, Center for Metals in Biocatalysis, and Supercom-
Fig. 1 Structurally defined 1 : 1 Cu/O2 adducts and the ligand used in
puter Institute, University of Minnesota, 207 Pleasant St. SE, Minneapolis,
this study (L, L1, and L2)
MN 55455-0431, USA. E-mail: tolman@chem.umn.edu; Fax: +1 612-624-
7029; Tel: +1 612-625-4061
Previous studies have shown that electronic influences of
ligand substituent variation on Cu(I)/O2 reactions can be sig-
nificant. Karlin and coworkers reported that electron-donating
4-methoxy or -amino substituents on the pyridyl donors of
poly(pyridyl)alkylamine ligands significantly influence the oxy-
genation kinetics of Cu(I) complexes, the ratio and spectroscopic
† The HTML version of this article has been enhanced with additional
colour images.
‡ Electronic supplementary information (ESI) available: NMR spectra, X-
ray structures of LCuCO, L2CuCO, and L2Cu(CH3CN), cyclic voltam-
mograms, UV-vis spectra, and selected results of theoretical calcula-
tions, and atomic coordinates for all computed structures. See DOI:
10.1039/b609939d
4944 | Dalton Trans., 2006, 4944–4953
This journal is
The Royal Society of Chemistry 2006
©

View Online
2
2
properties of the resulting l-g :g -peroxo and bis(l-oxo)dicopper
products, and their reactivity with added substrates.13,14 Some
of these effects have been examined by theory,15 and related
results were reported by Itoh and coworkers upon variation of
the alkylamine groups in similar ligands.16 Extensive data on the
electronic effects of CF3 substituents in tris(pyrazolyl)hydroborate
(Tp) complexes of Cu(I) have been published, with the m(CO)
values for TpCu(I)–CO compounds being particularly useful
indicators of the electron density at the metal.17 Gorun and
coworkers observed stabilization of a peroxodicopper complex
by a CF3-substituted Tp ligand, which appears to derive from
a combination of steric and electronic factors.18,19 Perhaps most
germane to our work is the study by Sadighi and coworkers
of Cu(I) complexes 5 (X = C6H6), in which incorporation of
the CF3 groups resulted in observation of arene substituent
hydroxylation upon room temperature oxygenation.20 Copper(I)
complexes of related fluorinated triazapentadienide ligands also
have been characterized, but no O2 reactivity was reported.21
Herein, we report a systematic study aimed at assessing elec-
tronic effects at parity of steric factors on the properties of Cu(I)
complexes of L (R = Rꢀ = Me, Fig. 1) and derivatives in which
the backbone methyl substituents are replaced by one (L1) or two
(L2) CF3 groups. The course of their low temperature oxygenations
are presented, along with theoretical calculations which provide
electronic and structural insights into the observed chemistry.
were carried out by bubbling dry O2 gas through the analyte
solution at ambient pressure. Excess O2 was removed by purging
the resulting solution with dry Ar gas. Samples for resonance
Raman spectroscopy were formed by creating a headspace of
16O2 over a solution of the Cu(I) complex in a Schlenk flask at
−80 ◦C or by adding ∼10 mL of 18O2 or 16O18O (50% 18O) to
a frozen solution of the Cu(I) complex (−196 ◦C) and reacting
the mixture at −80 ◦C. Solutions were prepared in acetone
with Cu(I) complex concentrations of 12.5 mM for the 16O2 and
18O2 reactions, and 18.0 mM for the 16O18O reaction. Resonance
Raman spectra were collected on an Acton AM-506 spectrometer
using a Princeton Instruments liquid N2 cooled (LN1100-PB)
CCD detector and ST-1385 controller interfaced with Winspec
software. A Spectra-Physics BeamLok 2060-KR-V Krypton ion
laser was used to excite at 413.1 nm. The spectra were obtained at
−196 ◦C using a backscattering geometry; samples were frozen in
a copper cup attached to a liquid nitrogen cooled coldfinger. The
spectralwindow was calibrated withliquidindene andspectra were
internally referenced to solvent. Baseline corrections (polynomial
fits) were carried out using Grams32/Spectral Notebase Version
4.04 (Galactic). Cyclic voltammograms were recorded using Pt
working and auxiliary electrodes, a Ag wire/AgNO3 (10 mM
in CH3CN) reference electrode, and a BAS Epsilon potentiostat
connected to a 22 mL cell in an inert-atmosphere glovebox.
Experiments were performed on analyte solutions of 1 mM in
THF with 0.3 M Bu4NPF6 (sample volumes of ∼5 mL) at room
temperature. The ferrocene/ferrocenium couple was recorded for
reference, using the reported value of E1/2 = + 560 mV vs. SCE
(for 0.1 M Bu4NPF6 in THF).25 FT-IR spectra were recorded
using a CaF2 solution cell (International Crystal Labs) in an
Avatar 370 spectrometer (ThermoNicolet). Elemental analyses
were performed by Robertson Microlit (Madison, NJ). ESI-MS
were recorded on a Bruker BioTOF II instrument.
Experimental
General considerations
All reagents were obtained from commercial sources and
used without further purification, unless otherwise stated.
Reagent grade 2,6-diisopropylaniline (90%) was freshly distilled
from CaH2. Deuterated benzene and THF were dried over
Na/benzophenone and vacuum transferred. Toluene, THF, pen-
tane and Et2O were passed through solvent purification columns
(Glass Contour, Laguna, CA). Acetonitrile was passed through
a solvent purification column (M-Braun) and further dried over
activated 3 A molecular sieves. Acetone (protio and deuterated
forms) was vacuum transferred from CaSO4 before drying over ac-
tivated 3 A molecular sieves (2–3 cycles). All metal complexes were
prepared and stored in a Vacuum Atmospheres inert atmosphere
glove-box under a dry nitrogen atmosphere or were manipulated
using standard inert atmosphere vacuum and Schlenk techniques.
The reagents (Me3SiCH2Cu)4,22 HL2,23 and LCu(CH3CN)24 were
prepared according to literature procedures. Labeled dioxygen
(18O2, and 50% statistical 18O mixed-label gas) was purchased from
Icon Isotopes, Inc.
X-Ray crystallography
Crystals of an appropriate size were placed onto the tip of a 0.1 mm
diameter glass capillary and mounted on either a Siemens or
Bruker SMART Platform CCD diffractometer.26 Data collections
were carried out using Mo Ka radiation at 173 K, with a detector
distance of ∼4.9 cm. A randomly oriented region of reciprocal
space was surveyed to the extent of one sphere and to a resolution
˚
˚
1
2
2
˚
˚
˚
of 0.77 A (L or L Cu(CH3CN)), 0.80 A (L Cu(CO)), and 0.84 A
(LCu(CO)). Four major sections of frames were collected with
0.30◦ steps in x at four different φ settings and a detector position
of −28◦ in 2h. The intensity data were corrected for absorption
and decay (SADABS).27 Final cell constants were calculated
from the xyz centroids of strong reflections from the actual data
collection after integration (SAINT).28 See ESI‡ for full crystal
and refinement details in the form of CIFs.
Physical methods
CCDC reference numbers 614585–617544.
NMR spectra were recorded on a Varian VI-500 or VI-300
spectrometer. NMR spectra were referenced internally to the
solvent (for benzene: 128.39 ppm for 13C NMR and 7.16 ppm in 1H
NMR for residual protio solvent). UV-Vis spectra were recorded
on an HP8453 (190–1100 nm) diode array spectrophotometer.
Low-temperature spectra were acquired using a Unisoku cryostat-
controlled UV-Vis cell holder. Solutions of Cu(I) complexes were
placed in anaerobic cuvettes sealed with rubber septa. Oygenations
For crystallographic data in CIF or other electronic format see
DOI: 10.1039/b609939d
HL1
In a 500 mL Schlenk flask, 2,6-diisopropylaniline (32 mL,
152.5 mmol) was combined with toluene (60 mL). The solution
was cooled in an ice–ethanol bath (≤ −10 ◦C) under a stream of
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4944–4953 | 4945
©

View Online
argon. A solution of TiCl4 (6.0 mL, 54.6 mmol) in toluene (10 mL)
was added by cannula over 15 min. The reaction mixture turned
brown and contained much insoluble material with about half
of the TiCl4/toluene mixture added. The reaction was gradually
warmed from −10 ◦C to 26 ◦C and stirred for 1 d. 1,1,1-Trifluoro-
2,4-pentanedione (3.0 mL, 24.8 mmol) was added dropwise to the
stirred mixture. With about a third of the dione added the reaction
turned yellow–orange and required addition of further toluene
(60 mL) to assist stirring. The reaction flask was fitted with a reflux
condenser and submerged in an oil bath. The mixture was heated at
88 ◦C (oil bath temperature) for 3 d, then at reflux for a further 2 d.
The cooled yellow sludgy-mixture was filtered under vacuum and
the red–brown filtrate washed with distilled water (3 × 100 mL).
Sol◦vents were removed on a Schlenk line using a water bath at
50 C. The red residue was purified by column chromatography
on silica gel using dichloromethane (Rf = 0.92). Recrystallisation
from hot methanol or ethanol afforded crystalline, yellow material.
Yield: 1.39 g (12%). 1H NMR (500 MHz, benzene-d6): d 1.07 (d,
6H, 3JHH = 6.5 Hz), 1.12 (apparent t—two overlapping doublets,
CF3), 123.26, 123.85, 124.10 and 124.73 (CH of phenyl rings),
140.01, 141.27, 147.39 and 148.14 (quaternary C of phenyl rings),
2
149.04 (q, JCF = 23.9 Hz, quaternary C of ligand backbone,
CCF3), 166.43 (quaternary C of ligand backbone, CCH3) ppm.
Anal. Calc. for C31H41CuF3N3: C, 64.62; H, 7.17; N, 7.29. Found:
C, 64.45; H, 7.40; N, 7.32.
1
L2Cu(CH3CN). Yield: 76%. H NMR (500 MHz, benzene-
d6): d − 0.02 (s, 3H, CH3CN), 1.34 (apparent t—two overlapping
doublets, 24H, 3JHH = 6.5 Hz, (CH3)2CH), 3.35 (sept, 4H, 3JHH
=
6.8 Hz, CH(CH3)2), 6.19 (s, 1H, CH of ligand backbone), 7.00–
7.10 (m, 6H, Ar–CH of phenyl rings) ppm. 13C NMR (75 MHz,
benzene-d6): d 0.03 (CH3CN), 22.95 and 25.49 ((CH3)2CH), 28.98
(CH(CH3)2), 83.94 (sept, 3JCF = 5.1 Hz, CH of ligand backbone),
1
121.68 (q, JCF = 286.1 Hz, CF3), 117.24 (CH3CN), 123.43 and
125.08 (CH of phenyl rings), 140.33 and 146.19 (quaternary C of
2
phenyl rings), 151.67 (q, JCF = 24.8 Hz, quaternary C of ligand
backbone) ppm. Anal. Calc. for C31H38CuF6N3: C, 59.08; H, 6.08;
N, 6.67. Found: C, 58.99; H, 5.85; N, 6.66.
12H, 3JHH = 7.3 Hz) and 1.31 (d, 6H, 3JHH = 7.5 Hz): (CH3)2CH,
(L, L1, or L2)Cu(CO). General procedure
3
1.51 (s, 3H, CH3 of ligand backbone), 3.11 (sept, 2H, JHH
=
3
The complex (L, L1, or L2)Cu(CH3CN) (29 lmol) was dissolved
in THF (2 mL) in a 10 mL Schlenk tube and stirred under
a CO headspace at ambient pressure for 2 h, during which
the solvents were evaporated. Half of the dried solid (pale
yellow to yellow) was dissolved in THF (0.4 mL) and the IR-
spectrum recorded immediately (CaF2); NMR analysis indicated
quantit◦ative conversion to the carbonyl adducts. After several days
at −20 C, pentane solutions of (L or L2)Cu(CO) produced yellow
crystals which were appropriate for X-ray structural analyses.
Elemental analyses were also performed on samples of these
crystals.
6.8 Hz) and 3.29 (sept, 2H, JHH = 6.8 Hz), CH(CH3)2, 5.40 (s,
1H, CH of ligand backbone), 7.07–7.18 (m, integration obscured
by protio-solvent peak, aromatic CH of phenyl rings), 12.12 (s,
1H, NH) ppm. 13C NMR (75 MHz, relax. delay = 8 s, benzene-
d6), d 21.12 (CH3 of ligand backbone), 23.30, 23.83, 24.79 and
3
25.52: (CH3)2CH, 28.96 and 29.24 (CH(CH3)2), 90.58 (q, JCF
=
3.8 Hz, CH of ligand backbone), 120.83 (q, 1JCF = 284.5 Hz, CF3),
123.78, 124.24, 126.23 and 127.56 (CH of phenyl rings), 138.74,
141.04, 141.75 and 143.60 (quaternary C of phenyl rings), 150.23
2
(q, JCF = 26.9 Hz, quaternary C of ligand backbone, CCF3),
164.02 (quaternary C of ligand backbone, CCH3) ppm. ESI-MS
(low resolution): m/z 473.3 [M + H+]. Anal. Calc. for C29H39F3N2:
C, 73.70; H, 8.32; N, 5.93. Found: C, 73.61; H, 8.60; N, 6.15.
LCu(CO). 1H NMR (500 MHz, benzene-d6): d 1.20 (d, 12H,
3JHH = 7.0 Hz) and 1.29 (d, 12H, 3JHH = 7.0 Hz): (CH3)2CH; 1.73
3
(L1 or L2)Cu(CH3CN). General procedure
(s, 6H, CH3 of ligand backbone), 3.35 (sept, 4H, JHH = 6.8 Hz,
CH(CH3)2), 4.98 (s, 1H, CH of ligand backbone), 7.09–7.15 (m,
6H, CH of phenyl rings) ppm. 13C NMR (125 MHz, benzene-d6):
d 23.31, 23.62, 24.88, 28.52, 95.49, 124.05, 125.44, 140.73, 149.35,
164.97, 178.42 ppm. FT-IR (THF): 2070 cm−1 (mCO). Anal. Calc.
for C30H41CuN2O: C, 70.76; H, 8.12; N, 5.50. Found: C, 70.96; H,
8.02; N, 5.39.
The ligand precursor HL1 or HL2 (200 mg) was dissolved in
CH3CN (10 mL) and added dropwise to (Me3SiCH2Cu)4 (1 mol.
eq. Cu) with stirring in a reaction vessel protected from light.
The reaction mixture was vigorously stirred for 1–2 d, the solvent
removed and the residue dissolved in CH3CN (∼2 mL). The
solution was filtered through a 0.2 micron filter disk before
refrigeration at −20 ◦C. Large crystals formed, and these were
separated from the mother liquor and washed with cold pentane.
The washings and mother liquor were combined, the solvent
removed and the residue dissolved in CH3CN in order to produce
further crop(s) of crystals upon refrigeration at −20 ◦C.
L1Cu(CO). 1H NMR (500 MHz, benzene-d6): d 1.12 (d, 6H,
3JHH = 7.0 Hz), 1.21 (d, 6H, 3JHH = 7.0 Hz), 1.26 (d, 6H, 3JHH
=
7.0 Hz) and 1.31 (d, 6H, 3JHH = 7.0 Hz): (CH3)2CH; 1.57 (s, 3H,
CH3 of ligand backbone), 3.09 (sept, 2H, 3JHH = 6.8 Hz) and 3.34
3
(sept, 2H, JHH = 6.8 Hz): CH(CH3)2; 5.55 (s, 1H, CH of ligand
backbone), 7.07–7.13 (m, 6H, CH of phenyl rings) ppm. 13C NMR
(125 MHz, benzene-d6): d 22.94, 23.38, 23.53, 24.72, 25.66, 28.59,
28.88, 91.00 (q, 3JCF = 4.9 Hz), 122.05 (q, 1JCF = 285.1 Hz), 123.68,
124.27, 125.86, 126.15, 139.77, 141.27, 147.50, 148.07, 150.32 (q,
2JCF = 24.3 Hz), 168.65, 176.39 ppm. FT-IR (THF): 2083 cm−1
(mCO).
1
L1Cu(CH3CN). . Yield: 97%. H NMR (500 MHz, benzene-
d6): d 0.00 (s, 3H, CH3CN), 1.23 (d, 6H, 3JHH = 7.0 Hz), 1.36 (d, 6H,
3JHH = 7.0 Hz) and 1.40 (apparent t—two overlapping doublets,
3
12H, JHH = 6.3 Hz): (CH3)2CH, 1.71 (s, 3H, CH3 of ligand
3
backbone), 3.35 (sept, 2H, JHH = 6.8 Hz) and 3.56 (sept, 2H,
3JHH = 6.8 Hz): (CH3)2CH, 5.56 (s, 1H, CH of ligand backbone),
7.03–7.15 (m, 6H, Ar–CH of phenyl rings) ppm. 13C NMR
(125 MHz, benzene-d6): d 0.23 (CH3CN), 23.19, 23.55, 23.84,
24.82 and 25.54 (all (CH3)2CH, and CH3 of ligand backbone),
L2Cu(CO). 1H NMR (500 MHz, benzene-d6): d 1.17 (d, 12H,
3
3JHH = 7.0 Hz) and 1.24 (d, 12H, JHH = 7.0 Hz): (CH3)2CH;
3
3.09 (sept, 4H, JHH = 6.8 Hz, CH(CH3)2), 6.18 (s, 1H, CH of
3
28.57 and 28.90 ((CH3)2CH), 89.16 (q, JCF = 4.5 Hz, CH of
ligand backbone), 7.05 (br s, 6H, CH of phenyl rings) ppm. 13C
NMR (125 MHz, benzene-d6): d 22.84, 25.55, 28.95, 85.73 (sept,
1
ligand backbone), 116.88 (CH3CN), 122.57 (q, JCF = 286.4 Hz,
4946 | Dalton Trans., 2006, 4944–4953
This journal is
The Royal Society of Chemistry 2006
©

View Online
3JCF = 5.1 Hz), 121.14 (q, 1JCF = 285.5 Hz), 123.89, 126.62, 140.18,
2
146.37, 153.52 (q, JCF = 25.6 Hz), 174.53 ppm. FT-IR (THF):
2097 cm−1 (mCO). Anal. Calc. for C30H35CuF6N2O: C, 58.38; H,
5.72; N, 4.54. Found: C, 58.32; H, 5.57; N, 4.39.
L1Cu(O2)
UV-Vis: A Beer–Lambert plot was constructed for replicate
oxygenations of L1Cu(CH3CN) in THF at −80 ◦C (0.2–0.6 mM);
[kmax, nm (e, M−1 cm−1)]: 415 (sh, 1780), 540 (125). Resonance
Raman (kex = 413.1 nm, 77 K, acetone): 977 (16O2), 955 (16O18O),
928 (18O2) cm−1.
Computational Methods: A. Density functional calculations
Scheme 1 Free energy cycles used to determine (a) pKb values and (b)
reduction potentials (see text).
Geometry optimizations were carried out using the Jaguar suite,
version 6.1, of ab initio quantum chemistry programs.29 Density
Functional Theory (DFT) with the B3LYP functional30–32 was
used in all calculations, as this functional has proven successful in
predicting ligand–Cu(I) and –Cu(II) bond dissociation energies.33
Computations on singlet and triplet states employed restricted
(RDFT) and unrestricted (UDFT) levels of theory, respectively,
in keeping with the established methodology for dealing with
complexes between Cu(I) and dioxygen.7b,8,11b,34,35 For copper, the
lacvp** effective core potential basis set36 was used, while 6-31G**
was used for all other atoms.
DGa = DGg − DGHsoAl + DG + DG
where DGg is the gas-phase deprotonation energy and DG , DG
(1)
A−
H+
sol
sol
HS
sol
A−
,
sol
and DGHso+l are the solvation free energies for the conjugate acid and
base forms and the proton, respectively. The solvation free energy
for the proton is taken to be −265.9 kcal mol−1 prior to correcting
for the 1 atm to 1 M standard-state change.44 The pKb is given by
eqn (2), where R is the universal gas constant and T is temperature.
DGa
2.303RT
pKb = 14.0 −
(2)
Relative electronic differences between various oxygenated
species were measured using Mulliken charge populations. Such
analyses, when founded on DFT calculations carried out with
balanced basis sets, have been effective at tracking electron flow in
a variety of reactions involving the reaction of dioxygen at a metal
center.7b,8,37,38
A free energy cycle was also used to calculate adiabatic reduction
potentials (Scheme 1b). DGr, the free energy change for reduction
in solution, is given by
X−
DGr = GXg − − GgX + DG − DGsXol
(3)
sol
where GXg − and GXg are gas phase free energies for the reduced and
oxidized species and DGsXo−l and DGsXol are their respective solvation
energies. The absolute reduction potential is computed by eqn (4),
Analytical vibrational frequency calculations were carried out
to verify optimized geometries as stationary points and to allow
for zero-point energy, enthalpy, and entropy corrections to be
made (and, hence, free energies to be computed). In order to
make these calculations more tractable, truncated models were
employed for L, L1, and L2 (Fig. 1), in which the four isopropyl
groups on the flanking 2,6-diisopropylphenyl rings were changed
to hydrogen atoms. The positions of these hydrogen atoms were
optimized with the remainder of the structure held fixed prior
to the frequency calculations. In subsequent calculations, the
full models were reoptimized with B3LYP, using the 6-311G**
triple-f plus polarization basis set for all atoms, except Cu, on
which the LACV3P** basis set (a triple-f basis compatible with the
effective core potential)36 was used. The O–O stretch frequencies
determined from vibrational frequency calculations on truncated
models derived from these triple-f geometries were used for
comparison with those determined spectroscopically,9 following
scaling by a factor of 0.97.39
DGr
nF
E◦ = −
(4)
where F is the Faraday constant and n is the number of
electrons transferred. Subtracting 4.28 V45 from E◦ in eqn (4)
yields the potential relative to the standard hydrogen electrode
(SHE). Systematic error in the computation of the reduction
potentials was accounted for using an empirical correction. Having
a copper coordination environment similar to that with the b-
diketiminate ligand and having been well characterized with
respect to its electrochemistry, a copper complex supported by
a triazamacrocyclic ligand was chosen as the reference system.46
Its measured Cu(III)/Cu(II) reduction potential was −0.271 V vs.
SHE, whereas the computed reduction potential was 0.485 V,
implying the correction factor to be −0.76 V. The viability of
transferring such corrections between closely related systems has
been established.47
Single-point solvation energies were calculated for each opti-
mized geometry using the self-consistent reaction field method as
implemented in the Poisson–Boltzmann solver in Jaguar.40,41 The
solvent was taken to be THF, in order to facilitate comparison
with experimental results, with dielectric constant e equal to 7.43
at 25 ◦C and 12.27 at −80 ◦C.42,43
B. Multireference calculations
Closed-shell singlet and high-spin triplet Kohn–Sham wave func-
tions can be expressed as single Slater determinants. However,
a minimum of two determinants is required to describe open-
shell singlets, which therefore cannot be rigorously expressed
within the framework of Kohn–Sham DFT. This shortcoming
A Born–Haber free energy cycle (Scheme 1a) was used to
determine pKb values. The free energy change for deprotonation
in aqueous solution was determined according to eqn (1),
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4944–4953 | 4947
©

View Online
of DFT is particularly relevant to calculations on complexes
between Cu(I) and dioxygen.48 These species typically possess
nontrivial degrees of Cu(II)–superoxo character, which may be
considered to be biradical, with one electron localized to Cu(II)
and the other to superoxide. In order to rigorously account
for the multideterminantal nature of the singlet 1 : 1 Cu/O2
adducts, single-point multireference second-order perturbation
theory (CASPT2) was used.49
In order to render the inherently expensive CASPT2 cal-
culations computationally feasible, procedure dictates chang-
ing both the backbone CH3/CF3 groups and flanking 2,6-
diisopropylphenyl groups of the b-diketiminate ligand to hydrogen
atoms prior to the CASPT2 calculations.7,8,11,35 Singlet–triplet
energy splittings are computed at the DFT and CASPT2 levels for
these small models, enabling the quantity D to be computed (eqn
(5)). With triplet states being well described by DFT, the relative
energy difference between the DFT and CASPT2 values for the
triplet energy can be taken as zero. D then becomes a measure
of the relative energy difference of the singlet states between the
two levels of theory and can be used as a correction to the DFT
energies for the singlet states. Final energies for the full model
system are obtained by combining the DFT energy for the full
model with the D value for the corresponding small model system.
In previous work examining the oxygenation of LCu(CH3CN)
(Fig. 1),7 we established that D linearly correlates with the Cu–
O bond length in the case of both side-on and end-on dioxygen
coordination to copper, and this result is used to expedite the
determination of D for the neutral singlet Cu/O2 adducts examined
here. For the protonated forms of the 1 : 1 adducts, single-point
multireference second-order perturbation theory (CASPT2) was
used.50 The initial complete active space (CAS) of the reference
wave function was comprised of 18 electrons in 12 orbitals
originating from the Cu valence orbitals and electrons and the
dioxygen r2p, r2p*, p2p, and p2p* orbitals and electrons. Eliminating
doubly occupied orbitals led to a final (12,9) active space. A 17-
electron relativistic effective core potential basis was used for Cu51
in these calculations, with a polarized double-f atomic natural
orbital basis48 being employed for all other atoms.
Scheme 2 Synthesis of ligands and complexes.
2,6-diisopropylaniline both in the double condensation and in the
removal of HCl from the reaction mixture, with the concomitant
formation of TiO2 providing the driving force. Upon addition
of the asymmetric ketone 1,1,1-trifluoro-2,4-pentanedione to the
TiCl4/diisopropylaniline mixture, prolonged reflux was required
before HL1 was detected by ESI-MS. Isolation of HL1 necessitated
the use of column chromatography and recrystallisation. Initial
unsuccessful efforts toward the synthesis of HL1 included a similar
reaction progression for the condensation, employing AlMe3 to
activate the 2,6-diisopropylaniline N-donor. The failure of this
approach is unsurprising given the low-yielding TiCl4-assisted
condensation.
The complexes (L1 or L2)Cu(CH3CN) were cleanly prepared
in high yields (97 and 76%, respectively) by direct reaction
of the corresponding ligands with the copper–alkyl reagent
[CuCH2Si(CH3)3]4 in acetonitrile (Scheme 2). In both cases,
crystals obtained from the reaction mixture were analytically pure
and suitable for X-ray structural determinations. Furthermore,
solids obtained by solvent evaporation from the reaction mixture
were free of contaminants (NMR), making this method convenient
1
for preparing Cu(I) complexes of high purity. The H and 13C
3
3
D = (1A − A)CASPT2 − (1A − A)DFT = [(1A)CASPT2
NMR spectra were fully assigned on the basis of splitting patterns,
coupling constants (e.g., JCF) and 2D correlations from HMQC
and HMBC experiments (Fig. S1, for example).‡ The complexes
(L, L1 or L2)Cu(CO) were formed by taking a THF solution
of the parent Cu(I)–acetonitrile complex to dryness under a
carbon monoxide atmosphere. The solids were characterized by
1H and 13C NMR and FT-IR, which indicated complete loss of
acetonitrile, with the only impurity being trace THF. Crystals of (L
or L2)Cu(CO) were further characterized, with elemental analyses
and X-ray structures (Fig. S2 and S3, respectively)‡ confirming
the presence of the terminal CO ligand. This series augments
the previously reported complexes 5 (X = CO), the only other
examples of Cu(I)–CO complexes supported by b-diketiminate
ligands.20
− (1A)DFT] − [(3A)CASPT2 − (3A)DFT
]
(5)
This procedure involving a combination of DFT and CASPT2
calculations has proven to reliably account for the multideter-
minantal nature of singlet 1 : 1 Cu/O2 adducts which occur in
biomimetic7 and active site modeling52 of DbM and PHM. The
methodology has yielded singlet–triplet state orderings consistent
with spectroscopic data for the adducts.7b,8,53 Experimental kinetics
data for the oxygenation of LCu(MeCN) was also well reproduced
using this computational methodology.7b
Results
Synthesis and structural characterization of copper(I) complexes
X-Ray
structural
determinations
were
performed
The complex LCu(CH3CN) was prepared from LiL and
[Cu(CH3CN)4](CF3SO3), as previously reported.24 HL1 was syn-
thesized in a similar manner to HL2 by a one-pot reaction, with
the first step being the activation of 2,6-diisopropylaniline by
coordination to titanium.23 The overall reaction (Scheme 2) utilizes
for L1Cu(CH3CN) (Fig. 2) and L2Cu(CH3CN) (Fig. S4).
‡The structures are very similar to that of LCu(CH3CN).24 The
Cu centers are trigonally coordinated by the three nitrogen donors,
with the bite-angle of the b-diketiminate (BDI) ligand (∼99◦)
somewhat smaller than the NBDI–Cu–Nnitrile angles (∼129–132◦)
4948 | Dalton Trans., 2006, 4944–4953
This journal is
The Royal Society of Chemistry 2006
©

View Online
◦
a
˚
Table 2 Properties of Cu(I) complexes of b-diketiminate ligands
Table 1 Selected bond distances (A) and angles ( )
LCu(CH3CN)b
L1Cu(CH3CN)
L2Cu(CH3CN)
Ligand
E
1/2/mVa
m(CO)/cm−1b
Bond or Angle
L
−96
110
411
2070
2083
2097
Cu–N1
Cu–N2
Cu–N3
1.9404(16)
1.9425(17)
1.864(2)
1.9431(14)
1.9307(16)
1.8656(16)
1.9398(14)
1.9352(14)
1.8698(16)
L1
L2
a Measured for Cu(I)–MeCN complex vs. Fc/Fc+. Conditions: 25 ◦C,
N1–Cu–N2
N1–Cu–N3
N2–Cu–N3
C2–N1–C6
C4–N2–C18
98.98(7)
131.75(8)
129.25(8)
118.92(17)
119.23(17)
99.04(6)
130.10(7)
130.78(7)
122.91(15)
119.34(16)
98.98(6)
128.93(7)
131.78(7)
124.74(14)
125.00(14)
Bu4NPF6 in THF (0.3 M). b Measured for Cu(I)–CO complex (THF).
Effects of CF3 groups on properties of Cu(I) complexes
a Estimated standard deviations are in parentheses. b The numbering
scheme for this structure is in accord with the CCDC entry (XIGWIT),
though differing slightly from the CIF accompanying ref. 24 (for which C6
and C18 appear as C11 and C31, respectively).
We assessed the electronic influences of the CF3 substituents
by comparing (a) the redox potentials of the complexes (L, L1
or L2)Cu(CH3CN) by cyclic voltammetry and (b) the m(CO)
values of the carbonyl complexes (L, L1, or L2)Cu(CO). Cyclic
voltammograms of (L, L1 or L2)Cu(CH3CN) measured in THF
with Bu4NPF6 (0.3 M) as electrolyte exhibited reversible waves
(ipc/ipa ∼ 1, DE = 89–116 mV at 20 mV s−1) that we attribute to
the Cu(I/II) couple (Fig. S5).‡54 The measured E1/2 values are
listed in Table 2. Substitution of one methyl group in L with
a CF3 moiety (L1) results in a ∼200 mV increase in the redox
potential, consistent with electronic destabilization of the Cu(II)
state by the electron withdrawing group.55 Substitution with a
second CF3 group (L2) results in a greater increase of ∼300 mV. The
E1/2 value is sufficiently high (∼+0.9 V vs. SCE) to suggest that
reactivity of L2Cu(CH3CN) with O2 would be thermodynamically
unfavorable, to the extent that electron transfer impacts complex
stability (see below).
The trend in m(CO) values for (L, L1, or L2)Cu(CO) is also
consistent with the expected effects of CF3 substitution, with
increases of ∼13 cm−1 for exchange of each methyl group by a
CF3 unit. These increases reflect greater C–O bond order due
to decreased electron back donation from the less electron rich
Cu(I) center. Similar trends have been described for TpCu(I)CO
complexes (Tp = tris(pyrazolyl)hydroborate), for which the effects
of extensive substituent variations on m(CO) values have been
mapped.17 While the b-diketiminate ligand is generally considered
to be more electron donating than Tp,56 it is noteworthy that the
m(CO) value for L2Cu(CO) of 2097 cm−1 is essentially identical
to that for the Cu(I)–CO complex of the Tp ligand having a
CF3 substituent at each pyrazolyl 3-position (2100 cm−1). Finally,
further support for the correlation of m(CO) with the electron
withdrawing effects of the CF3 groups is provided by the trend in
the carbonyl carbon chemical shifts in 13C NMR spectra, which
are 178.42, 176.39, and 174.53, respectively, for the complexes
supported by L, L1, and L2.
Fig. 2 Representation of the X-ray structure of L1Cu(CH3CN), showing
all nonhydrogen atoms as 50% thermal ellipsoids. Unlabeled atoms are
carbons.
(Table 1). The steric environments provided by L1 and L2 are
essentially the same as that of L, as gauged by the non-perturbed
Ca–NBDI–Caryl angles (∼119–125◦, Table 1, last two entries). The
CF3 substituent(s) are accommodated in the ligand backbone
without a collapse of the aryl ring(s) toward the acetonitrile
ligand. Thus, L, L1 and L2 represent a progression of ligands
that differ in their electron donating capabilites (vide infra) while
maintaining similar, demanding steric environments at the Cu(I)
site.
DFT calculations were performed for the Cu(I) complexes, and
good agreement between the optimized geometries and crystal
structures was observed. Heavy-atom RMSD values between
˚
˚
˚
the two sets of structures are 0.028 A, 0.040 A, and 0.045 A
with L,7 L1, and L2, respectively. In the calculated structures,
Cu–Nnitrile distances increase from 1.934 A (L) to 1.970 A (L )
to 1.972 A (L ) as the electron-donating character of the b-
O2 Reactivity of Cu(I) complexes: A. Experiment
1
˚
˚
Solutions of (L1 or L2)Cu(CH3CN) in THF were oxygenated at
−80 ◦C, but only in the case of L1Cu(CH3CN) was an intermediate
observed spectroscopically. No new features in the UV-vis spec-
trum were observed upon extensive oxygenation of L2Cu(CH3CN)
2
˚
diketiminate ligand is decreased by addition of the CF3 groups.
This trend is not possible to discern experimentally (Table 1)
because the differences are the same order of magnitude as the
estimated standard deviations in the interatomic distances. As with
the Cu(I)–CH3CN complex supported by the anilido-imine ligand
in 4,8 no significant asymmetry occurs with respect to CH3CN
◦
between −80 and 25 C. For L1Cu(CH3CN), UV-vis features at
415 (sh, e = 1780 M−1cm−1) and 540 (e = 125 M−1 cm−1) nm
appeared upon treatment with O2, which decayed upon warming
coordination in L1Cu(CH3CN), with a difference in the two NBDI
Cu–Nnitrile bond angles of only 0.68◦.
–
(Fig. S6).‡ The resonance Raman spectrum acquired using kex
=
413.1 nm (77 K, acetone) exhibited a peak at 977 cm−1 which
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4944–4953 | 4949
©

View Online
Table 3 Spectroscopic data for Selected Cu/O2 Complexes
Complex
UV-vis (kmax (e))
m(O–O) (16O2/16O18O/18O2)
Ref.
3 (R = Me)
3 (R = tBu)
4
385 (2400), 600 (200)
424 (2000), 638 (200)
390 (7600), 645 (250)
415 (1780), 540 (125)
968/943/917
961/937/912
974/—/908
977/955/928
7
7
8
L1CuO2
This work
2
1
2
shifted to 928 cm−1 when 18O2 was used (Fig. 3). When a mixture
of 18O2, 16O18O, and 16O2 was used, the spectrum contained an
additional band at 955 cm−1. The UV-vis and resonance Raman
data are similar to those reported for 3 and 4 (Table 3), and on this
basis we assign the oxygenation product as an analogous side-on
adduct L1CuO2. The m(O–O) values for L1CuO2 are higher than
those of 3 (R = Me or tBu) as would be expected for decreased
reduction of the O2 ligand in the system supported by the poorer
electron donor L1. However, the differences are small, indicating
a relatively minor influence of the CF3 substituent on the bonding
in the CuO2 unit.
Table 4 Calculated free energy differences (g –g ) between side-on (g )
and end-on (g ) singlet Cu/O2 adducts supported by L, L1, and L2 and
1
free energy changes (DG) for the oxygenation reactiona
2
1
Complex
g –g
DG
LCuO2
L1CuO2
L2CuO2
−4.7b
−2.4
−0.4
−9.9b
−5.1
+0.4
a At −80 ◦C in THF, except as noted. All energies in kcal mol−1
THF (ref. 7b).
.
−50 C,
◦
b
structure for a precatalytic complex of PHM4 and complex 2,6
is nearly isoenergetic with the side-on isomer. The free energy
of oxygenation (−80 ◦C, THF) concurrently increments by
∼5 kcal mol−1 with the addition of each CF3 group (Table 4),
such that oxygenation of the Cu(I)–CH3CN complex, which was
exergonic by 9.9 kcal mol−1 in the case of L, becomes slightly
endergonic by 0.4 kcal mol−1 in the case of L2. This marked
increase is consistent with the inability to experimentally observe
oxygenation of L2Cu(MeCN). With regards to synthesis of an end-
on 1 : 1 Cu/O2 adduct, a challenging reality is thus revealed here.
Use of a less electron donating ligand (such as L2) favors end-on
dioxygen coordination, but copper complexes supported by such
ligands are also markedly more difficult to oxygenate.
The O–O bond distances in the end-on adducts decrease by
1
˚
˚
0.006 A with the addition of each CF3 group from 1.281 A in g -
1
2
˚
LCuO2 to 1.269 A in g -L CuO2. Similarly, in the side-on cases, the
˚
O–O bond lengths shorten by 0.005 A per CF3 group leading from
Fig. 3 Resonance Raman spectra of the product of oxygenation
of L1Cu(CH3CN) using 16O2 (solid line), 18O2 (dotted line), and a mixture
of 16O2, 16O18O, and 18O2 (dashed line). The peak positions of the O-isotope
sensitive features are indicated.
2
2
2
˚
˚
an O–O distance in g -LCuO2 of 1.358 A to 1.348 A in g -L CuO2.
For L1Cu(CH3CN), the asymmetry of the ligand backbone does
not translate into asymmetric dioxygen binding. The NBDI–Cu–
Oproximal angles in g -L1CuO2 differ by only 0.2 ; the Cu–O distances
in g -L CuO2, by only 0.004 A. Such virtually symmetric dioxygen
coordination was also observed in the case with the asymmetric
anilido-imine ligand, 4.8
◦
1
2
1
˚
B. Theory
In addition to the trend in O–O bond distances between L, L1,
and L2, computed O–O stretch frequencies increase by 8–9 cm−1 in
both the end-on and side-on (L, L1 or L2)CuO2 complexes for each
CF3 group substituted into the ligand backbone (Table 5). The
magnitude of this change is close to that measured experimentally
For each ligand system and both end-on and side-on dioxygen
coordination, singlet states are predicted by theory to be lower in
energy than the corresponding triplet states by 2–10 kcal mol−1
(Table S1).‡ In the case of end-on dioxygen binding to copper, the
singlet–triplet splitting remains fairly constant across the series of
ligands. On the other hand, the triplet states become relatively
more stable in the case of side-on O2 coordination as Cu(II)–
superoxide character increases with increasing ligand fluorination
(vide infra). All subsequent discussion will focus on the singlet
states of (L, L1 or L2)CuO2 given their energetic stability over the
triplet states.
1
2
Table 5 Computed m(O–O) values for singlet end-on (g ) and side-on (g )
Cu/O2 adductsa
1
2
Complex
g (16O2/16O18O/18O2)
g (16O2/16O18O/18O2)
LCuO2
L1CuO2
L2CuO2
1238/1198/1165b
1247/1216/1175
1256/1222/1185
1042/1012/983b
1050/1022/991
1061/1032/1001
2
1
The 4.7 kcal mol−1 preference for the g versus g dioxygen
coordination in LCuO2 decreases by ∼2 kcal mol−1 for each
7b
CF3 group substituted onto the ligand backbone (Table 4).
With L2, end-on O2 coordination to copper, as seen in the crystal
a Units: cm−1 b Ref. 7b.
.
4950 | Dalton Trans., 2006, 4944–4953
This journal is
The Royal Society of Chemistry 2006
©

View Online
Table 6 Mulliken charge populations on the dioxygen fragment in singlet
Table 7 Calculated reduction potentials and pKb values for the singlet
side-on Cu/O2 adducts and gas phase bond dissociation energies (BDEs)
for a hydrogen atom bound to the dioxygen fragment
1
2
end-on (g ) and side-on (g ) Cu/O2 adducts
1
2
Complex
g
g
a
b
Complex
E◦
pKb
BDEc
LCuO2
L1CuO2
L2CuO2
−0.43a
−0.33
−0.32
−0.60a
−0.52
−0.50
LCuO2
L1CuO2
L2CuO2
−2.06d
−1.81
−1.56
14.1d
17.5
23.4
54.9
56.8
58.7
a Ref. 7b.
a Units: V vs. SHE, at −80 ◦C in THF. b At 25 ◦C in H2O. c Units: kcal mol−1
.
d Ref. 11b.
between L1CuO2 and 3 (R = Me; Table 3).7 Comparison of the
computed m(16O2) and the differences Dm(18O2) = m(16O2) − m(18O2)
1
2
for g - and g -L1CuO2 supports the assignment of the exper-
imental product obtained from oxygenation of L1Cu(CH3CN)
as the side-on 1 : 1 adduct (cf. Table 3).ठExperimentally non-
resolvable m(O–O) for the mixed label isotopomers (Cu–16O18O
vs. Cu–18O16O) were found in the end-on cases, providing more
examples where mixed label resonance Raman experiments are not
diagnostic for the mode of coordination of dioxygen to the metal
center.6a,57
Decomposing hydrogen atom abstraction into its
components60—transfer of a proton and transfer of an electron
from substrate to the 1 : 1 adduct—provides insight into the origin
of the lack of reactivity with the 1 : 1 Cu/O2 species (for details,
see ESI‡ text and Tables S2–S5).‡ Low reduction potentials
and high pKb values for the 1 : 1 Cu/O2 adducts (Tables 7, S6)‡
indicate significant stability toward reduction or protonation.11b
These same data reveal that, as the electron donating capability
of the ligand is decreased by incrementing the number of CF3
backbone groups, the reduction potentials and pKb values both
increase. The former trend in particular mirrors that seen in the
experimental redox potentials for the Cu(I)–MeCN complexes,
which also increased by 200–300 mV upon substitution of each
CF3 group (cf. Table 2). In sum, reduction becomes promoted
while protonation is deterred, and these effects offset one
another to lead to minimal change in the O–H BDEs, and hence
reactivity, among the Cu/O2 adducts supported by each of L, L1,
and L2.
Mulliken charge population analysis (Table 6) confirms that
the degree of dioxygen reduction decreases in going from LCuO2
to L1CuO2 to L2CuO2 in both the end-on and side-on cases.
This in turn is fully consistent with the aforementioned decreases
in O–O bond length and increases in m(O–O) which occurs
upon CF3 substitution. These three trends combined indicate
2
that while g -LCuO2 lies far to the Cu(III)–peroxide side7,10 of
the Cu(II)–superoxide/Cu(III)–peroxide continuum,9 the side-on
oxygen complexes supported by L1 and L2 each possess more
Cu(II)–superoxide than the former. The narrow range of these
trends, however, indicates that the overall electronic differences
2
2
2
between g -LCuO2, g -L1CuO2, and g -L2CuO2 are relatively
small and thus the inductive effect of the CF3 groups, while
unmistakable, is nonetheless minor.
Conclusions
Copper(I)–acetonitrile complexes with the ligands L1 and L2, and
the series (L, L1 or L2)Cu(CO) were prepared and characterized.
Along with the previously reported complex LCu(CH3CN), these
copper(I) species provided the means for an examination of the
potential to electronically tune the b-diketiminate ligand and
thereby influence the outcomes of Cu(I)/O2 reactivity. As such, the
donating ability of the ligands was assessed through comparisons
of the redox potentials of the nitrile adducts and the carbonyl
stretching frequencies of the carbon-monoxide adducts. The shift
in potential for the Cu(I/II) redox pair in going from L to L1 to L2
(spanning ∼500 mV) reflects the electronic destabilisation of the
Cu(II) state—a direct consequence of the lessened ligand electron
donation upon CF3 substitution. For L2Cu(CH3CN), the redox
potential alone is suggestive of a ligand environment not conducive
to O2 reactivity. The increase in m(CO) with increasing ligand
backbone fluorination is substantial and the trend correlates with
the redox potential data.
While L1Cu(CH3CN) undergoes rapid oxygenation with O2, L2
does not support the formation of a Cu/O2 adduct. L1Cu(O2)
was formulated as a side-on bound Cu(III)–peroxo complex on the
basis of electronic and vibrational spectroscopy, including simi-
larities with LCu(O2), and theoretical predictions that coincided
with the observed oxygen-isotope sensitivity of the O–O vibration.
Although the presence of one CF3 substituent (L1) had a small
effect on the Cu/O2 chemistry relative to the case for L (as gauged
by the slight shift in m(O–O) for its 1 : 1 adduct), the electronics
Finally, the effects of CF3 substitution on the reactivity of the
Cu/O2 adducts were assessed by theoretical evaluation of the
capability of the adducts to abstract H atoms from hydrocarbons.
This can be measured by determining the O–H bond dissociation
energies (BDEs) for the O–H bonds formed upon hydrogen atom
abstraction. With the side-on Cu/O2 adducts, the BDEs rise
from 54.9 kcal mol−1 to 58.7 kcal mol−1 in going from L to L2
(Table 7). The O–H BDEs for the end-on adducts are relatively
insensitive to changes in the ligand backbone at ∼60 kcal mol−1
(Table S6).‡ Comparison to C–H BDEs such as those for 1,4-
cyclohexadiene (73.0 kcal mol−1)58 and the benzylic position in the
physiological substrate for DbM,59 dopamine (87.2 kcal mol−1),11b
implies that the 1 : 1 adducts supported by each of L, L1, and L2
should be unreactive towards substrates. Despite the observed
∼4 kcal mol−1 increase in potency for the side-on adducts
upon CF3 substitution into the ligand backbone, their ability to
attack substrate nonetheless remains well below that predicted for
Cu(III)–oxo species.11b,12d,e
§ DFT calculations have systematically overestimated O–O stretching
frequencies in 1 : 1 Cu/O2 adducts.7b,8,53 This is most likely attributable to
the challenge of obtaining a sufficiently accurate wave function for these
multideterminantal species within the framework of DFT (see computa-
tional methods).34 While quantitative comparison to experimental m(O–
O) may therefore be problematic, qualitative comparisons to experimental
m(O–O) values and analysis of the relative computed m(O–O) values provide
useful insights into the nature of the Cu/O2 species.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4944–4953 | 4951
©

View Online
are clearly perturbed in such a way that the presence of two CF3
substituents (L2) prevents formation of a Cu/O2 intermediate.
Further, the theoretical comparisons for oxygenation of (L, L1
or L2)Cu(CH3CN) indicate that the formation of an end-on 1 : 1
Cu/O2 adduct requires a less-donating b-diketiminate ligand (such
as L2), though this in itself proves to be an experimental obstacle
given that such ligands give rise to an endergonic oxygenation
process. Thus, future efforts in developing a bidentate ligand
that supports an end-on 1 : 1 Cu/O2 adduct face the challenge
of balancing suitable ligand electronics with experimentally fea-
sible (thermodynamically favorable) oxygenation of the Cu(I)
complex.
13 (a) C. X. Zhang, S. Kaderli, M. Costas, E. Kim, Y.-M. Neuhold, K. D.
Karlin and A. D. Zuberbu¨hler, Inorg. Chem., 2003, 42, 1807; (b) J.
Shearer, C. X. Zhang, L. N. Zakharov, A. L. Rheingold and K. D.
Karlin, J. Am. Chem. Soc., 2005, 127, 5469.
14 M. J. Henson, M. A. Vance, C. X. Zhang, H.-C. Liang, K. D. Karlin
and E. I. Solomon, J. Am. Chem. Soc., 2003, 125, 5186.
15 A. J. Johansson, M. R. A. Blomberg and P. E. M. Siegbahn, Inorg.
Chem., 2006, 45, 1491.
16 T. Osako, S. Terada, T. Tosha, S. Nagatomo, H. Furutachi, S. Fujinami,
T. Kitagawa, M. Suzukib and S. Itoh, Dalton Trans., 2005, 3514.
17 H. V. R. Dias and T. K. H. H. Goh, Polyhedron, 2004, 23, 273, and
references cited therein.
18 (a) Z. Hu, R. D. Williams, D. Tran, T. G. Spiro and S. M. Gorun, J. Am.
Chem. Soc., 2000, 122, 3556; (b) Z. Hu, G. N. George and S. M. Gorun,
Inorg. Chem., 2001, 40, 4812.
19 G. Aullo´n, S. M. Gorun and S. Alvarez, Inorg. Chem., 2006, 45, 3594.
20 D. S. Laitar, C. J. N. Mathison, W. M. Davis and J. P. Sadighi, Inorg.
Chem., 2003, 42, 7354.
21 H. V. R. Dias and S. Singh, Inorg. Chem., 2004, 43, 5786.
22 J. A. J. Jarvis, R. Pearce and M. F. Lappert, J. Chem. Soc., Dalton
Trans., 1977, 999.
23 D. T. Carey, E. K. Cope-Eatough, E. Vilaplana-Mafe´, F. S. Mair, R. G.
Pritchard, J. E. Warren and R. J. Woods, Dalton Trans., 2003, 1083.
24 D. J. E. Spencer, N. W. Aboelella, A. M. Reynolds, P. L. Holland and
W. B. Tolman, J. Am. Chem. Soc., 2002, 124, 2108.
25 N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877.
26 SMART v5.054, Bruker Analytical X-Ray Systems, Madison, WI,
2001.
Acknowledgements
We thank the NIH (Grant GM47365 to W. B. T., Grant NRSA-
GM070144 to B. F. G.) and the NSF (Grant CHE-0203346 to
C. J. C.) for financial support. We also thank Dr Victor G. Young,
Jr. for assistance with X-ray data collection and refinement, and
Professor Lawrence Que, Jr. for the use of resonance Raman spec-
troscopy instrumentation. Elizabeth Devine and Nicole Settergren
are thanked for initial experimental efforts. We are also grateful to
the reviewers for their helpful suggestions.
27 An empirical correction for absorption anisotropy: R. Blessing, Acta
Crystallogr., Sect. A., 1995, A51, 33.
28 For L2Cu(CH3CN): SAINT v6.2, Bruker Analytical X-Ray Systems,
Madison, WI, 2001; For L1Cu(CH3CN), LCu(CO) and L2Cu(CO):
SAINT+ v6.45, Bruker Analytical X-Ray Systems, Madison, WI,
2003.
References
29 Jaguar 6.1. Schrodinger, LLC: Portland, Oregon, 2003.
30 B. G. Johnson, P. M. W. Gill and J. A. Pople, J. Chem. Phys., 1993, 98,
5612.
1 (a) M. A. Halcrow, in Comprehensive Coordination Chemistry II, ed.
J. A. McCleverty and T. J. Meyer, Elsevier: Amsterdam, 2004, vol. 8,
pp. 395; (b) S. Itoh, Curr. Opin. Chem. Biol., 2006, 10, 115.
2 I. Arends, P. Gamez and R. A. Sheldon, Adv. Inorg. Chem., 2006, 58,
235.
3 (a) L. M. Mirica, X. Ottenwaelder and T. D. P. Stack, Chem. Rev., 2004,
104, 1013; (b) E. A. Lewis and W. B. Tolman, Chem. Rev., 2004, 104,
1047.
4 S. T. Prigge, B. A. Eipper, R. E. Mains and L. M. Amzel, Science, 2004,
304, 864.
5 (a) K. Fujisawa, M. Tanaka, Y. Moro-oka and N. Kitajima, J. Am.
Chem. Soc., 1994, 116, 12079; (b) P. Chen, D. E. Root, C. Campochiaro,
K. Fujisawa and E. I. Solomon, J. Am. Chem. Soc., 2003, 125,
466.
31 A. D. Becke, J. Chem. Phys., 1993, 98, 1372.
32 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785.
33 J.-M. Ducere, A. Goursot and D. Berthomieu, J. Phys. Chem. A, 2005,
109, 400.
34 B. F. Gherman and C. J. Cramer, Inorg. Chem., 2004, 43, 7281.
35 N. W. Aboelella, B. F. Gherman, L. M. R. Hill, J. T. York, N. Holm,
V. G. Young, Jr., C. J. Cramer and W. B. Tolman, J. Am. Chem. Soc.,
2006, 128, 3345.
36 (a) P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270; (b) W. R.
Wadt and P. J. Hay, J. Chem. Phys., 1985, 82, 284; (c) P. J. Hay and
W. R . Wa d t , J. Chem. Phys., 1985, 82, 299.
37 R. A. Friesner, M.-H. Baik, V. Guallar, B. F. Gherman, M. Wirstam,
R. B. Murphy and S. J. Lippard, Coord. Chem. Rev., 2003, 238–239,
267.
6 (a) M. Schatz, V. Raab, S. P. Foxon, G. Brehm, S. Schneider, Markus
Reiher, M. C. Holthausen, J. Sundermeyer and S. Schindler, Angew.
Chem., Int. Ed., 2004, 43, 4360; (b) C. Wu¨rtele, E. Gaoutchenova, K.
Harms, M. C. Holthausen, J. Sundermeyer and S. Schindler, Angew.
Chem., Int. Ed., 2006, 45, 3867.
7 (a) N. W. Aboelella, E. A. Lewis, A. M. Reynolds, W. W. Brennessel,
C. J. Cramer and W. B. Tolman, J. Am. Chem. Soc., 2002, 124, 10660;
(b) N. W. Aboelella, S. V. Kryatov, B. F. Gherman, W. W. Brennessel,
V. G. Young, Jr., R. Sarangi, E. V. Rybak-Akimova, K. O. Hodgson, B.
Hedman, E. I. Solomon, C. J. Cramer and W. B. Tolman, J. Am. Chem.
Soc., 2004, 126, 16896.
38 M.-H. Baik, M. Newcomb, R. A. Friesner and S. J. Lippard, Chem.
Rev., 2003, 103, 2385.
39 C. W. Bauschlicher and H. Partridge, J. Chem. Phys., 1995, 103,
1788.
40 B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N.
Ringnalda, D. Sitkoff and B. Honig, J. Phys. Chem., 1996, 100, 11775.
41 D. J. Tannor, B. Marten, R. B. Murphy, R. A. Friesner, D. Sitkoff, A.
Nicholls, M. N. Ringnalda, W. A. Goddard, III and B. Honig, J. Am.
Chem. Soc., 1994, 116, 11875.
42 C. Carvajal, K. J. Tolle, J. Smid and M. Szwarc, J. Am. Chem. Soc.,
1965, 87, 5548.
43 D. J. Metz and A. Glines, J. Phys. Chem., 1967, 71, 1158.
44 M. D. Tissandier, K. A. Cowen, W. Y. Feng, E. Gundlach, M. H. Cohen,
A. D. Earhart, J. V. Coe and T. R. Tuttle, J. Phys. Chem. A, 1998, 102,
7787.
8 A. M. Reynolds, B. F. Gherman, C. J. Cramer and W. B. Tolman, Inorg.
Chem., 2005, 44, 6989.
9 C. J. Cramer, W. B. Tolman, K. H. Theopold and A. L. Rheingold,
Proc. Natl. Acad. Sci. USA, 2003, 100, 3635.
10 R. Sarangi, N. Aboelella, K. Fujisawa, W. B. Tolman, B. Hedman, K. O.
Hodgson and E. I. Solomon, J. Am. Chem. Soc., 2006, 128, 8286.
11 (a) A. M. Reynolds, E. L. Lewis, N. W. Aboelella and W. B. Tolman,
Chem. Commun., 2005, 2014; (b) B. F. Gherman, W. B. Tolman and
C. J. Cramer, J. Comput. Chem., 2006, in press.
12 (a) J. P. Evans, K. Ahn and J. P. Klinman, J. Biol. Chem., 2003, 278,
49691; (b) P. Chen and E. I. Solomon, J. Am. Chem. Soc., 2004, 126,
4991; (c) B. F. Gherman, D. E. Heppner, W. B. Tolman and C. J. Cramer,
J. Biol. Inorg. Chem., 2006, 11, 197; (d) T. Kamachi, N. Kihara, Y. Shiota
and K. Yoshizawa, Inorg. Chem., 2005, 44, 4226; (e) K. Yoshizawa,
N. Kihara, T. Kamachi and Y. Shiota, Inorg. Chem., 2006, 45,
3034.
45 C. P. Kelly, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2006,
110, 16066.
46 X. Ribas, D. A. Jackson, B. Donnadieu, J. Mahia, T. Parella, R. Xifra,
B. Hedman, K. O. Hodgson, A. Llobet and T. D. P. Stack, Angew.
Chem., Int. Ed., 2002, 41, 2991.
47 P. Winget, C. J. Cramer and D. G. Truhlar, Theor. Chem. Acc., 2004,
112, 217.
48 K. Pierloot, B. Dumez, P. O. Widmark and B. O. Roos, Theor. Chim.
Acta, 1995, 90, 87.
4952 | Dalton Trans., 2006, 4944–4953
This journal is
The Royal Society of Chemistry 2006
©

View Online
49 K. Andersson, P. A. Malmqvist and B. O. Roos, J. Chem. Phys., 1992,
96, 1218.
50 G. Karlstrom, R. Lindh, P. A. Malmqvist, B. O. Roos, U. Ryde, V.
Veryazov, P. O. Widmark, M. Cossi, B. Schimmelpfennig, P. Neogrady
and L. Seijo, Comput. Mater. Sci., 2003, 28, 222.
51 Z. Barandiaran and L. Seijo, Can. J. Chem., 1992, 70, 409.
52 B. F. Gherman, D. E. Heppner, W. B. Tolman and C. J. Cramer, J. Biol.
Inorg. Chem., 2006, 11, 197.
Holland, B. A. Jazdzewski, C. Duboc-Toia, L. L. Pape, S. Yokota, Y.
Tachi, S. Itoh and W. B. Tolman, Inorg. Chem., 2002, 41, 6307.
55 D. B. Rorabacher, Chem. Rev., 2004, 104, 651.
56 D. W. Randall, S. DeBeer, P. L. Holland, B. Hedman, K. O. Hodgson,
W. B. Tolman and E. I. Solomon, J. Am. Chem. Soc., 2000, 122,
11632.
57 C. R. Kinsinger, B. F. Gherman, L. Gagliardi and C. J. Cramer, J. Biol.
Inorg. Chem., 2005, 10, 778.
53 N. W. Aboelella, B. F. Gherman, L. M. R. Hill, J. T. York, N. Holm,
V. G. Young, Jr., C. J. Cramer and W. B. Tolman, J. Am. Chem. Soc.,
2006, 128, 3445.
58 D. R. Lide, Handbook of Chemistry and Physics, 85th Edition, CRC
Press, Boca Raton, 2004.
59 (a) J. P. Klinman, Chem. Rev., 1996, 96, 2541; (b) L. C. Stewart and J. P.
Klinman, Annu. Rev. Biochem., 1988, 57, 551.
60 J. M. Mayer, Acc. Chem. Res., 1998, 31, 441.
54 A value of E1/2 = −215 mV vs. Fc/Fc+ was reported previously for
LCu(CH3CN) in CH3CN: D. J. E. Spencer, A. M. Reynolds, P. L.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4944–4953 | 4953
©