Polyhedron 30 (2011) 178–186

Contents lists available at ScienceDirect

Polyhedron
journal homepage: www.elsevier.com/locate/poly

Chiral copper–bipyridine complexes: Synthesis, characterization

and mechanistic studies on asymmetric cyclopropanation
Wing-Sze Lee a, Chi-Tung Yeung a, Kiu-Chor Sham a, Wing-Tak Wong b, Hoi-Lun Kwong a,⇑
a
Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong SAR, China
b
Department of Chemistry, The University of Hong Kong, Pokfulam Road, Hong Kong SAR, China

a r t i c l e i n f o a b s t r a c t

Article history: Chiral bipyridine ligands of different steric properties when reacted with CuCl2 formed orange, yellow or
Received 21 July 2010 green solids of new copper(II) complexes, [Cu(L)Cl2] (L = L2–6), in good yield. Together with [Cu(L1)Cl2],
Accepted 12 October 2010 these complexes were characterized in solution by UV–Vis spectroscopy and cyclic voltammetry. The
Available online 5 November 2010
complexes give d–d transitions between 860 and 970 nm, and exhibit one quasi-reversible Cu(II)/Cu(I)
couple between +0.405 V and +0.516 V versus NHE. Two of the copper(II) complexes, [Cu(L5)Cl2] and
Keywords: [Cu(L6)Cl2], and a copper(I) complex of L1, [Cu(L1)Cl], were characterized by X-ray crystallography.
Chiral bipyridines
The triflate derivatives of both the Cu(I) and Cu(II) complexes are active catalysts towards the cycloprop-
Copper–bipyridine complexes
Asymmetric catalysis
anation of ethyl diazoacetate with styrene. The asymmetric induction suffers when the size difference
Cyclopropanation between the alkyl and alkoxyl groups was minimized. The mechanism of the cyclopropanation was stud-
ied with kinetic and competition experiments. The rate is first order in catalyst and ethyl diazoacetate,
inverse order with styrene and is strongly affected by the counterion.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction plexes with ligands having different steric properties. Herein, we

report the synthesis, spectroscopic characterisation and redox
Copper–diimine complexes have been of great interest in a behaviours of a series of copper(II)–bipyridine complexes, as well
number of research areas, such as photochemistry [1,2], supramo- as their use in the asymmetric cyclopropanation of styrene with
lecular chemistry [3,4] electrochemistry [5,6] and catalysis [7–9]. ethyl diazoacetate (EDA). X-ray crystal structures for [Cu(L5)Cl2]
In enantioselective reactions, chiral versions of the complexes have and [Cu(L6)Cl2] and a copper(I) complex, [Cu(L1)Cl], are described.
also been used successfully in catalysis that involved cyclopropa- The results of mechanistic studies and kinetic studies under
nation [10–13] and allylic oxidation [14–16]. The results led us pseudo-first order conditions on the asymmetric cyclopropanation
to systematically investigate a family of chiral copper(II) com- catalysts are also presented.

N N N N N N

OMe MeO O O OH HO
PhCH2 CH2Ph
L1 L2 L3

N N N N N N

OMe MeO O O OH HO
PhCH2 CH2Ph
L4 L5 L6

⇑ Corresponding author. Tel.: +852 27887304; fax: +852 27887406.

E-mail address: bhhoik@cityu.edu.hk (H.-L. Kwong).

0277-5387/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2010.10.021
 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186 179

13
2. Experimental C NMR (CDCl3): d 22.7, 70.9, 78.8, 78.9, 119.7, 127.5, 127.6,
128.3, 137.5, 138.3, 155.3, 162.6. Anal. Calc. for C28H28O2N2: C,
2.1. General methods 79.22; H, 6.65; N, 6.60. Found: C, 79.29; H, 6.54; N, 6.34%.

Toluene was distilled under N2 over sodium. Dichloromethane 2.3. Procedure for the preparation of [Cu(L)Cl2]
and acetonitrile were distilled over calcium hydride. Diethyl ether
and THF were distilled under N2 over sodium/benzophenone. Chiral bipyridine L (0.4 mmol) in CH2Cl2 (5 ml) was added drop-
Chemicals were of reagent-grade quality and were obtained com- by-drop to a solution of CuCl22H2O (0.4 mmol, 0.068 g) in absolute
mercially. Infrared spectra in the range 500–4000 cm1 using a ethanol (5 ml). The mixture was refluxed for a few hours. Addition
Nujol matrix or KBr plates were recorded on a Perkin–Elmer Model of ether to the cooled reaction mixture led to the formation of a
FTIR-1600 spectrometer. The electronic absorption spectra were microcrystalline solid. The mixture was placed in the refrigerator
measured on a Perkin–Elmer Lambda 19 double-beam UV–Vis- overnight and the microcrystalline solid was filtered and washed
NIR spectrophotometer. 1H and 13C NMR spectra were recorded with ether.
on a Varian 300 MHz Mercury instrument. Positive ion FAB mass
spectra as a 3-nitrobenzylalcohol matrix were recorded on a Finna- 2.3.1. [Cu(L2)Cl2]
gin MAT 95 spectrometer. ESI-MS were taken by a PE SCIEX API Recrystallization from CH2Cl2/EtOH/Et2O gave 0.136 g (53%) of
365 mass spectrometer. Electron ionization mass spectra were re- an orange solid: Anal. Calc. for CuN2Cl2C34H40O2: C, 63.49; H,
corded on a Hewlett–Packard 5890II GC instrument coupled with a 6.22; N, 4.36. Found: C, 64.10; H, 6.12; N, 4.48%. UV–Vis-NIR spec-
5970 mass selective detector. Elemental analyses were performed trum (CH2Cl2), kmax (nm) (e/M1 cm1): 245 (16 000), 311 (15 000),
on a Vario EL elemental analyzer. Optical rotations were measured 390 sh (852), 887 (150); MS (+FAB): 607(M+Cl) and 572
by a JASCO DIP-370 digital polarimeter. Melting points were mea- (M+2Cl).
sured by an electrothermal digital apparatus. The chiral bromopyr-
idine intermediates and bipyridines L1, L3 and L6 were prepared 2.3.2. [Cu(L3)Cl2]
according to the literature procedures [17]. [Cu(L1)Cl2] and Recrystallization from CH2Cl2/EtOH/Et2O gave 0.152 g (83%) of a
[Cu(L1)Cl] were synthesized according to our previously reported green solid: IR (KBr, cm1): 3268 s; Anal. Calc. for CuN2Cl2-
procedure [11]. C20H28O2(H2O): C, 49.94; H, 6.24; N, 5.83. Found: C, 49.98; H,
6.27; N, 5.79%. UV–Vis-NIR spectrum (MeOH), kmax (nm) (e/
2.2. Synthesis of bipyridines L2, L4 and L5 M1 cm1): 304 (11 400), 246 (10 700), 257 (9300), 871 (70); MS
(API): 390 (M+HClCl).
At 70 °C and under N2, to NiCl26H2O (6 mmol, 1.43 g) in de-
gassed DMF (30 ml), triphenylphosphine (24 mmol, 6.30 g) was 2.3.3. [Cu(L4)Cl2]
added to give a blue solution. Zinc powder (13 mmol, 0.87 g) was Recrystallization from CH2Cl2/EtOH/Et2O gave 0.109 g (67%) of a
then added and the resulting mixture was stirred for 1 h, resulting yellow solid: Anal. Calc. for CuN2Cl2C16H20O2: C, 47.23; H, 4.92; N,
in the formation of a dark-brown mixture. The suitable bromopyr- 6.89. Found: 47.57; H, 4.83; N, 7.01%. UV–Vis-NIR spectrum
idine (5 mmol) in degassed DMF (5 ml) was added slowly and the (CH2Cl2), kmax (nm) (e/M1 cm1): 315 (15 400), 301 (14 500),
mixture was stirred for another 3 h. The mixture was then allowed 246 (12 600), 366 sh (770), 996 (110); MS (+FAB): 370 (M+Cl)
to cool to room temperature and 5% aqueous NH3 (50 ml) was and 335 (M+2Cl).
added. The layers were separated, and the aqueous layer was ex-
tracted three times with CH2Cl2 (70 ml  3). The combined organic 2.3.4. [Cu(L5)Cl2]
layers were washed three times with water (50 ml  3) and once Recrystallization from CH2Cl2/EtOH/Et2O gave 0.132 g (59%) of a
with brine (50 ml). Drying with Na2SO4 and removal of the solvent yellow solid: Anal. Calc. for CuN2Cl2C28H28O2: C, 60.16; H, 5.01; N,
under reduced pressure yielded a pale yellow solid. This was puri- 5.01. Found: 61.13; H, 4.91; N, 5.11%. UV–Vis-NIR spectrum
fied by column chromatography (petroleum ether–ethyl acetate) (CH2Cl2), kmax (nm) (e/M1 cm1): 307 (16 000), 317 (14 900),
to give a white solid. 246 sh (11 000), 377 sh (738), 980 (130); MS (+FAB): 522
(M+Cl) and 587 (M+2Cl).

2.2.1. Bipyridine L2
2.3.5. [Cu(L6)Cl2]
Yield: 0.66 g (52%); 1H NMR (CDCl3): d 1.02 (s, 18H), 4.33 (s, 2H),
Recrystallization from MeOH/Et2O gave 0.064 g (42%) of a green
4.32–4.50 (m, 4H), 7.27–7.35 (m, 10H), 7.47 (d, 2H, J = 7.5 Hz), 7.81
solid: IR (KBr, cm1): 3534 versus; Anal. Calc. for CuN2Cl2C14H16O2:
(t, 2H, J = 7.5 Hz), 8.29 (d, 2H, J = 7.5 Hz); 13C NMR (CDCl3): d 26.3,
C, 44.39; H, 4.23; N, 7.40. Found: 44.44; H, 4.15; N, 7.49%. UV–Vis-
26.3, 26.4, 35.7, 71.3, 90.4, 119.4, 121.6, 121.6, 127.1, 128.0, 136.5,
NIR spectrum (MeOH), kmax (nm) (e/M1 cm1): 308 (10 300), 258
138.7, 154.8, 159.6. Anal. Calc. for C34H40O2N2: C, 80.28; H, 7.93; N,
(12 000), 318 (7900), 438 (100), 860 (70); MS (+FAB): 342 (M+Cl).
5.51. Found: C, 80.48; H, 7.70; N, 5.24%.
2.4. X-ray crystallography analysis for [Cu(L5)Cl2], [Cu(L6)Cl2] and
2.2.2. Bipyridine L4 [Cu(L1)Cl]
Yield: 0.38 g (56%); 1H NMR (CDCl3): d 1.53 (d, 6H, J = 6.6 Hz),
3.36 (s, 6H), 4.50–4.56 (m, 2H), 7.43 (d, 2H, J = 7.8 Hz), 7.83 (t, For [Cu(L5)Cl2] and [Cu(L1)Cl], diffraction data were obtained
2H, J = 7.8 Hz), 8.33 (d, 2H, J = 7.5 Hz); 13C NMR (CDCl3): d 22.3, on a Rigaku AFC7R diffractometer at 296 and 301 K, respectively.
56.8, 80.8, 119.5, 119.6, 137.3, 155.3, 162.2. Anal. Calc. for Absorption corrections based on the PSI scans technique were ap-
C16H20O2N2: C, 70.56; H, 7.40; N, 10.29. Found: C, 70.70; H, 7.38; plied on both of these complexes. The structures were solved by
N, 9.99%. using direct methods (SHELXS97) and refined on F2 against all reflec-
tions. The absolute configurations of [Cu(L1)Cl] at C11 and C17
2.2.3. Bipyridine L5 were found to be R, as confirmed by the Flack parameter of
Yield: 0.52 g (49%); 1H NMR (CDCl3): d 1.59 (d, 6H, J = 6.6 Hz), 0.000(14). For [Cu(L6)Cl2], diffraction data were obtained on a
4.47–4.58 (m, 4H), 4.72–4.79 (m, 2H), 7.30–7.37 (m, 10H), 7.54 Burker SMART 1000 CCD diffractometer at 298 K. The multi-scan
(d, 2H, J = 7.5 Hz), 7.84 (t, 2H, J = 7.8 Hz), 8.35 (d, 2H, J = 7.5 Hz); method was applied for absorption correction. The structure was
180 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186

Table 1
Crystallographic data for [Cu(L5)Cl2], [Cu(L6)Cl2] and [Cu(L1)Cl].

Complex [Cu(L5)Cl2] [Cu(L6)Cl2] [Cu(L1)Cl]

Formula C28H28Cl2CuN2O2 C14H16Cl2CuN2O2 C22H32ClCuN2O2
Molecular weight 558.96 378.73 455.49
Crystal color, habit yellow, rod green, block brown, rod
Crystal dimensions (mm) 0.12  0.07  0.07 0.22  0.22  0.10 0.07  0.21  0.23
Crystal system orthorhombic trigonal monoclinic
Lattice
a (Å) 12.947(3) 10.965(1) 9.7263(13)
b (Å) 25.011(3) 10.965(1) 11.3273(14)
c (Å) 8.044(4) 22.247(2) 11.5636(15)
a (°) 90 90.00 90
b (°) 90 90.00 112.837(9)
c (°) 90 120.00 90
V (Å3) 2604.7(15) 2316.4(4) 1174.1(3)
Space group P212121 (no. 19) P3121 (no. 144) P21
Z value 4 6 2
Dcalc (g/cm3) 1.425 1.629 1.288
T (K) 301 298 296(2)
Radiation, k (Å) Mo Ka, 0.71073 Mo Ka, 0.71073 Mo Ka, 0.71073
l (Mo Ka) (mm1) 1.07 1.76 1.06
2hmax (°) 50.0 55.0 45.0
F(0 0 0) 1156 1158 480
Reflections: measured/independent/with I > 2r(I) 2224/2224/689 14385/3490/2682 3482/3078/2718
Rint 0.000 0.040 0.022
R-factor: R1; wR 0.046; 0.166 0.032; 0.070 0.033; 0.090
Flack parameter – 0.000(14) 0.000(14)

solved by using the heavy atoms Patterson method (PATTY) and re- column. For competition experiments, an equal amount of styrene
fined on F2 against all reflections. The absolute configurations of (2 mmol) and substituted styrene (2 mmol) was used.
the molecule were found to be S for C2 and C13, as confirmed by
the Flack parameter of 0.000(14). Crystal data and details of the
2.7. Kinetic experiments (EDA decomposition)
measurements for these complexes are summarized in Table 1.

All the kinetics experiments reported here were carried out by

2.5. Cyclic voltammetry analysis
the following procedure: to a two-neck round bottom flask were
added [Cu(L1)Cl] (0.0046 g, 0.01 mmol), degassed chloroform
Cyclic voltammetry was performed using a CH Instruments
(5.0 ml) and silver(I) salt (0.01 mmol) under nitrogen. The solution
Electrochemical Workstation CHI750A. Experiments were carried
was stirred at room temperature for 10 min and filtered through a
out under nitrogen in a three-electrode cell with glassy-carbon as
packed filter paper to another flask containing a solution of chloro-
the working electrode, a Pt-wire as the counter electrode and sat-
benzene, an internal standard (0.01 M) in degassed chloroform
urated Ag/AgNO3 as the reference electrode. Ferrocene was used as
(95 ml). EDA (0.01 mmol) was then added and the reaction was
the internal standard and 0.1 M tetrabutylammonium hexafluoro-
monitored by GC until no EDA was present. For cyclopropanation
phosphate as a supporting electrolyte. Cyclic voltammograms were
with styrene, styrene was added together with the internal stan-
scanned at room temperature in the potential range between
dard. For the investigation of the counterion, suitable silver salts
+0.6 V and 0.2 V at the sweep rate of 0.1 V/s.
were used.

2.6. Procedure for copper-catalyzed cyclopropanation

3. Results and discussion
Under nitrogen, [Cu(L)Cl2] (0.02 mmol), AgOTf (0.04 mmol) and
CH2Cl2 (2 ml) were added to a two-neck round bottom flask and 3.1. Ligands and complexes syntheses
stirred at room temperature for 2 h. The mixture was filtered
through a packed filter-paper and styrene (4 mmol) was added. The chiral bipyridines L1–6, having different substituent
The catalyst was activated by the addition of 0.2 equiv of diazoace- groups, were prepared either as reported previously (L1, L3 and
tate. Some of the complexes required heating during the activation L6) or following a similar synthetic route (L2, L4 and L5) [17].
process. Using a syringe pump, a solution of diazoacetate (1 mmol) The copper(II) complexes [Cu(L)Cl2] (L = L1–6) were obtained in
in CH2Cl2 (0.5 ml) was added to the reaction mixture over a period moderate to good yields by the reaction of an equal molar ratio
of 4 h. After the addition of diazoacetate, the mixture was allowed mixture of L and CuCl22H2O in dichloromethane–ethanol solution
to stir for 16 h at room temperature. The solvent was then removed at reflux temperature. All the complexes were air-stable in the so-
and the crude product obtained was purified by column chroma- lid state and in solution. Elemental analyses indicate that these
tography (hexane/EtOAc). All the cyclopropanes obtained are compounds are 1:1 ligand-to-copper complexes with two chloride
known compounds and were characterized by 1H NMR, 13C NMR, ions. The positive ion FAB mass spectra for these complexes show
IR and GC–MS. The enantiomeric excesses of the cyclopropanes parent ion peaks with masses that matched with the formulas cal-
were determined by HPLC with a Daicel Chiralcel OJ column. Abso- culated by CHN analysis. The colors of the complexes are different:
lute configurations were determined by comparing the order of [Cu(L1)Cl2] and [Cu(L2)Cl2] are orange; [Cu(L4)Cl2] and [Cu(L5)Cl2]
elution with samples of a known configuration [18]. Diastereose- are yellow; and [Cu(L3)Cl2] and [Cu(L6)Cl2] are green. Both the
lectivities (cis/trans ratio) were measured by GC with an Ultra orange and yellow colors are unusual for Cu(II) complexes [19].
2-crosslinked 5% PhMesilcone (25 m  0.2 mm  0.33 lm) Copper(I) complexes of these ligands have also been prepared
 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186 181

R1 N N R2
L1-6 + CuCl2
CH2Cl2/ethanol R2 Cu R1
reflux OR RO
Cl Cl

[Cu(L1)Cl2]: R1 = t-Bu, R2 = H, R = CH3

[Cu(L2)Cl2]: R1 = t-Bu, R2 = H, R = CH2Ph
[Cu(L3)Cl2]: R1 = t-Bu, R2 = H, R = H
[Cu(L4)Cl2]: R1 = H, R2 = CH3, R = CH3
[Cu(L5)Cl2]: R1 = H, R2 = CH3, R = CH2Ph
[Cu(L6)Cl2]: R1 = H, R2 = CH3, R = H

Scheme 1. Synthesis of [Cu(L)Cl2].

0.8 L1
L2
0.7 0.16 L3
L4
0.6 L5
0.12 L6
Absorbance

0.5
0.08
A

0.4
0.04
0.3
0.00
0.2 600 800 1000 1200
nm
0.1

0.0
400 600 800 1000
Wavelength (nm)

Fig. 1. Electronic absorption spectra of (a) [Cu(L)Cl2] (L = L1–6) at a concentration

of 3.8  105 M at room temperature (CH2Cl2 was used for L1, L2, L4 and L5, MeOH
was used for L3 and L6). The inset shows spectra in the region 600–1200 nm at a
concentration of 1.0  103 M.

but these are mostly unstable, except for [Cu(L1)Cl] which is a red-
dish brown solid (Scheme 1). Fig. 2. Cyclic voltammograms of [Cu(L)Cl2] (L = L1–6 from top to bottom) in
acetonitrile with a scan rate of 0.1 V/s. All the potentials were measured against the
Ag/AgNO3 reference electrode.
3.2. Spectroscopic properties

As the color of the complexes [Cu(L)Cl2] vary from orange Table 2

(L = L1 and L2) to yellow (L = L4 and L5) to green (L = L3 and L6), Cyclic voltammetric parameters for [Cu(L)Cl2].
their spectroscopic behaviors were explored. UV–Vis-NIR elec-
Complex E1/2 (V) DEp (mV)a ip,c/ip,ab
tronic absorption spectra of [Cu(L)Cl2] (L = L1–6) are shown in
[Cu(L1)Cl2] +0.482 155 1.03
Fig. 1. Intense absorption bands at about 252 and 310 nm in the
[Cu(L2)Cl2] +0.516 134 0.49
UV-region may be attributed to p–p* transitions of the bpy ligands [Cu(L3)Cl2] +0.462 107 1.04
[20]. These absorption bands extend to the visible region with dif- [Cu(L4)Cl2] +0.405 166 0.83
ferent molar absorptivities. For the complexes with L1 and L5, rel- [Cu(L5)Cl2] +0.424 103 1.05
atively strong absorptions at 406 (e  800 M1 cm1) and 377 nm [Cu(L6)Cl2] +0.436 121 1.08

(e  700 M1 cm1) were observed respectively. These absorptions a

Cathodic peak to anodic peak separation.
are generally believed to originate from LMCT (Cl ? CuII) [20]. For b
Ratio of cathodic peak current to anodic peak current.
[Cu(L6)Cl2], the absorption for this transition is weak (438 nm,
e  100 M1 cm1). In addition to the above, all the complexes ex-
hibit a very broad band in the lower energy region (>600 nm), ten- (vide infra). In addition, the absorption at 980 nm for [Cu(L5)Cl2]
2
tatively assigned to a Cu(II) d–d transition [20]. Many studies have may be assigned to the transition of dxz or dyz ? dz and the absorp-
shown that the energies of these d–d transitions correlate with the tion at 860 nm for [Cu(L6)Cl2] may be assigned to the transition of
2 2
copper coordination geometries [21] and that the intensity maxi- dz ? dx  y2 [21,24]. For [Cu(L1)Cl2], as reported previously [11],
mum is shifted to lower energy when the geometry is changed the absorption at 919 nm indicates a pseudo-tetrahedral geometry
from a regular square pyramid to a regular trigonal bipyramid [25]. Complex [Cu(L2)Cl2], which has a color similar to [Cu(L1)Cl2],
[21–23]. This was also observed in our study, as [Cu(L6)Cl2], which has an absorption at 887 nm and might also have a tetrahedral
is a distorted square pyramid with s = 0.28 (vide infra), absorbed at geometry. When compared with [Cu(L6)Cl2] and [Cu(L5)Cl2],
a higher energy than [Cu(L5)Cl2] (860 versus 980 nm), which has respectively, the similar absorptions of [Cu(L3)Cl2] and [Cu(L4)Cl2]
comparatively more of a trigonal bipyramidal form with s = 0.48 possibly indicate similarities in geometries.
182 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186

Fig. 3. Crystal structures for (a) [Cu(L5)Cl2], (b) [Cu(L6)Cl2] and (c) [Cu(L1)Cl] including the atom numbering schemes. All hydrogen atoms have been omitted for clarity.

3.3. Redox behaviour and as the cathodic to anodic peak current ratios, ip,c/ip,a, deviate
from unity, all complexes exhibit quasi-reversible electron transfer
The redox behaviour of the Cu(II/I) couple in acetonitrile was behavior.
assessed by cyclic voltammetry. For comparison, the cyclic voltam- The E1/2 values for the complexes with ligands that contain the
mograms are shown in Fig. 2 and data summarized in Table 2. As bulkier t-butyl substituents (L1–3) are higher than the copper
the cathodic and anodic peak separations (DEp = 103  166 mV) complexes with the less bulky corresponding methyl substituents
are larger than those of a fully reversible process (DEp = 59 mV), (L4–6). This observation provides evidence that the steric
 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186 183

Table 3 the s index of 0.48 for [Cu(L5)Cl2] reflects the distorted nature of its
Selected bond distances and angles for [Cu(L5)Cl2], [Cu(L6)Cl2] and [Cu(L1)Cl]. complex.
[Cu(L5)Cl2] [Cu(L6)Cl2] [Cu(L1)Cl] [Cu(L6)Cl2] also has a distorted trigonal bipyramidal geometry
Atoms similar to [Cu(L5)Cl2]: one of the nitrogen atoms N(1) of the bipyr-
Bond lengths (Å) idine ligand and two chloride ions occupy the equatorial plane
Cu(1)–Cl(1) 2.311(5) 2.274(1) 2.129(3) while the oxygen atom O(1) and the remaining nitrogen atom
Cu(1)–Cl(2) 2.302(5) 2.424(1) – N(2) are axial. The copper–nitrogen bond distances (Cu(1)–
Cu(1)–N(1) 2.009(13) 1.959(3) 2.067(6)
Cu(1)–N(2) 2.065(14) 2.064(3) 2.054(10)
N(1) = 1.959(3) Å; Cu(1)–N(2) = 2.064(3) Å) are slightly shorter
Cu(1)–O(1) 2.135(13) 2.095(2) – than those of [Cu(L5)Cl2]. The two copper–chloride bonds are dif-
Bond angles (°)
ferent (Cu(1)–Cl(1) = 2.274(1) Å and Cu(1)–Cl(2) = 2.424(1) Å).
Cl(1)–Cu(1)–Cl(2) 118.86(19) 107.99(3) – The copper–oxygen bond is 2.095(2) Å. The bite angle of 80.0°
Cl(1)–Cu(1)–O(1) 95.97(37) 93.27(7) – made by the bipyridine ligand and copper is comparable to that
Cl(1)–Cu(1)–N(1) 114.18(144) 141.28(9) 138.7(1) of [Cu(L5)Cl2]. The O(1)–Cu(1)–N(2) angle of 157.4° (Table 2)
Cl(1)–Cu(1)–N(2) 95.52(40) 100.37(8) 140.69(9)
shows the O–Cu–N linkage is not perfectly linear. The bond angles
Cl(2)–Cu(1)–O(1) 89.75(40) 92.13(6) –
Cl(2)–Cu(1)–N(1) 126.23(39) 109.98(9) – for Cl(1)–Cu(1)–Cl(2), Cl(1)–Cu(1)–N(1) and Cl(2)–Cu(1)–N(1) are
Cl(2)–Cu(1)–N(2) 103.37(40) 100.55(8) – 108.0°, 141.3° and 110.0°, respectively. As the angles around the
O(1)–Cu(1)–N(1) 76.44(54) 78.11(12) – copper sum up to 359.3°, the equatorial plane of the complex is
O(1)–Cu(1)–N(2) 155.58(52) 157.41(10) – almost planar. Given its s index of 0.28, the structure of [Cu(L6)Cl2]
N(1)–Cu(1)–N(2) 79.19(57) 80.00(11) 80.61(12)
is more distorted and closer to a square pyramid.
Dihedral angles (°) The copper(I) centre of [Cu(L1)Cl] is three-coordinated and has
N(1)–C(x)–C(y)–N(2) 9.75(232) 6.14(42) 6.04(48)
x = 7, y = 8 x = 7, y = 8 x = 5, y = 6
a Y-shaped planar geometry, with the metal ion surrounded by two
pyridine nitrogen atoms and one chloride anion. The copper–
nitrogen bond distances are 2.054(10) and 2.067(6) Å. These bond
distances are slightly longer than the copper–nitrogen bond
bulkiness of L1–3 distorts the copper(II) state and hence makes the distances for other achiral three-coordinate copper(I) phenanthro-
complexes easier to reduce than those with ligands that are less line complexes. The bite angle of 80.6° made by the bipyridine li-
bulky. Among the three structurally characterized complexes (vide gand and copper is small when compared with the value of 82.9°
infra), the E1/2 for the tetrahedral complex, [Cu(L1)Cl2], is higher cited in the literature [31]. The bond angles for Cl(1)–Cu(1)–N(1)
than that for the square pyramid complex, [Cu(L6)Cl2]. The E1/2 va- and Cl(1)–Cu(1)–N(2) are 138.7° and 140.7°, respectively. The an-
lue for the trigonal bipyramidal complex, [Cu(L5)Cl2], is the lowest. gles around copper sum up to be 360° which suggests that the
The high E1/2 value for [Cu(L1)Cl2] is most probably due to the min- complex is planar in geometry. The copper–chloride bond distance
imal alteration to the preferred tetrahedral geometry when the is 2.129 Å. There are a few examples of three-coordinate Y-shaped
complex is reduced to Cu(I) [23,26,27]. Along the same line, in copper(I) complexes reported in the literature [32–34]. The large
the case of the two five-coordinated complexes, the higher E1/2 va- difference in geometry between [Cu(I)(L1)Cl] and [Cu(I)(L1)Cl2]
lue of [Cu(L6)Cl2] can be attributed to be the smaller change in its probably explains the quasi-reversible nature of the Cu(II)/Cu(I) re-
square pyramidal geometry during reduction from Cu(II) to Cu(I), dox couple.
whereas the lower E1/2 value of [Cu(L5)Cl2] likely reflects a signif-
icant change in geometry upon reduction. 3.5. Catalytic asymmetric cyclopropanation

3.4. Structural characterization The use of [Cu(L)Cl2] (L = L2–6) in the cyclopropanation of

styrene with EDA was studied and the results are summarized in
The crystal structures of [Cu(L5)Cl2], [Cu(L6)Cl2] and [Cu(L1)Cl] Table 4. The active form of the catalyst was generated in situ by
are shown in Fig. 3a–c. Selected bond lengths and angles are reacting 2 mol% of [Cu(L)Cl2] with AgOTf, which was then filtered
listed in Table 3. The copper(II) centres of both [Cu(L5)Cl2] and and reacted with a few equivalents of EDA to reduce the copper(II)
[Cu(L6)Cl2] are five-coordinated and have a distorted trigonal bipy- species to copper(I). Copper catalysts generated from L1–5 re-
ramidal geometry: one of the nitrogen atoms N(1) of the bipyridine quired no heating and were much more easily activated than the
ligand and two chloride ions occupy the equatorial plane while the one from L6, which required heating at 40 °C for a few minutes.
oxygen O(1) atom and the remaining nitrogen atom N(2) are axial. As L6 has the lowest E1/2 value, the result is consistent with our
The marked difference between these two structures and the pre- observation on the correlation between the Cu(II)/Cu(I) redox po-
viously reported [Cu(L1)Cl2] [11] demonstrate that the coordina- tential of the copper(II) chloride complex and the E1/2 value.
tion environment is affected by the steric property of the ligand. All these copper–bipyridine complexes were active catalysts,
In the structure of [Cu(L5)Cl2], the copper–nitrogen bond dis- producing cyclopropyl esters in yields between 85% and 99% (en-
tances for the two nitrogen atoms are different (Cu(1)– tries 1–6). The best results in both enantioselectivity and diastere-
N(1) = 2.009(13) Å, Cu(1)–N(2) = 2.065(14) Å) but they are within oselectivity (trans:cis = 80:20) were achieved with L1, which gave
the range generally found for other copper(II) bipyridine com- 90% ee and 82% ee for trans and cis-isomers, respectively (entry
plexes [28,29]. The copper–chloride bonds are 2.311(5) and 1). Trans/cis ratios ranged from 66:34 to 80:20 (entries 1–6). When
2.302(5) Å and the copper–oxygen bond is 2.135(13) Å. The bite L2 and L3 were employed, the absolute configurations of the cyclo-
angle of 79.2° made by the bipyridine ligand and copper is small propane esters were found to be (1R, 2R) and (1R, 2S) for the trans
when compared with that of 83.0° for [Cu(L1)Cl2] [11]. As the and cis isomers, respectively, similar to L1. Employment of L4–6 re-
O(1)–Cu(1)–N(2) angle is 155.6°, the O–Cu–N linkage for the com- sulted in the opposite absolute configurations (trans = 1S, 2S;
plex is not perfectly linear. The bond angles around the Cu(1) cen- cis = 1S, 2R). This can be attributed to the use of ligands with oppo-
tre Cl(1)–Cu(1)–Cl(2), Cl(1)–Cu(1)–N(1) and Cl(2)–Cu(1)–N(1) are site configuration, thus the sense of asymmetric induction ob-
118.9°, 114.2° and 126.2°, respectively. As the sum of these angles served is also consistent with L1 [11]. In general, better
is 359.3°, the equatorial plane of the complex is almost planar. As enantioselectivity and trans/cis ratio are observed with the t-butyl
the degree of distortion in the geometry of a five-coordinated com- series of ligands. Decreasing the size difference between the alkyl
plex can be represented by the value of its trigonality s index [30], group and the alkoxy/hydroxy group has a negative effect. The
184 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186

Table 4
Asymmetric cyclopropanation of styrene with ethyl diazoacetate, catalyzed by [Cu(L)(OTf)2].

O 2 mol% Cu(OTf)2
2.2 mol% L Ph H H H
+ H +
OEt
CH2Cl2 r.t., N2 H COOEt Ph COOEt
N2

Entrya Ligand Yield (%)b trans/cis % ee. (config.)c

trans cis
1 L1 85 80:20 90 (1R, 2R) 82 (1R, 2S)
2 L2 86 76:24 83 (1R, 2R) 57 (1R, 2S)
3 L3 92 74:26 73 (1R, 2R) 38 (1R, 2S)
4 L4 88 66:34 30 (1S, 2S) 25 (1S, 2R)
5 L5 92 66:34 43 (1S, 2S) 23 (1S, 2R)
6 L6 99 70:30 44 (1S, 2S) 26 (1S, 2R)
a
Diazoacetate (0.2 equivalent) was used for the reduction of the Cu(II) catalysts before the start of the cyclopropanation reaction.
b
Isolated yield after column chromatography.
c
Enantiomeric excesses were determined by HPLC with a Daicel Chiralcel OJ column, and the absolute configurations were determined by comparing the order of elution of
samples with a known configuration [18].

Table 5
Catalytic asymmetric cyclopropanation with [Cu(L1)X] as the catalyst.

O 1mol% [Cu(L)Cl]
H 1 mol% AgX Ph H H H
+ OEt +
CHCl 3 , r.t., N 2 H COOEt Ph COOEt
N2

Entry AgX Solvent trans:cis % ee (trans)a % ee (cis)a Yield %b

1 CHCl3 80:20 90 (1R, 2R) 82 (1R, 2S) 67
2 AgOTf CH2Cl2 79:21 91 (1R, 2R) 85 (1R, 2S) 68
3 Toluene 83:17 92 (1R, 2R) 78 (1R, 2S) 70
4 THF 82:18 93 (1R, 2R) 85 (1R, 2S) 76
5 CH3CN – – – –
6 AgPF6 CHCl3 75:25 86 (1R, 2R) 85 (1R, 2S) 88
7 THF 75:25 79 (1R, 2R) 70 (1R, 2S) 54
8 AgBF4 CHCl3 71:29 87 (1R, 2R) 83 (1R, 2S) 65
9 THF 75:25 91 (1R, 2R) 84 (1R, 2S) 74
10 AgSbF6 CHCl3 74:26 85 (1R, 2R) 83 (1R, 2S) 56
11 THF 76:24 90 (1R, 2R) 86 (1R, 2S) 43
12 AgClO4 CHCl3 – – – –
a
Enantiomeric excesses were determined by HPLC with a Daicel Chiralcel OJ column, and the absolute configurations were determined by comparing the order of elution of
samples with a known configuration [18].
b
Isolated yield after chromatography.

group on the oxygen has a relatively small effect on the enantiose- Cl. This is probably due to the strong coordination of the perchlo-
lectivity, probably because of the flexibility of the group and its rate ion to the metal center.
distance away from the metal centre. The role of the counterion was further investigated by a compe-
tition experiment with substituted styrenes. With [Cu(L1)X]
3.6. Effect of anion and competition experiments (X = OTf, PF6, BF4 and SbF6 as the catalyst, competitions were
carried out in either chloroform or THF. In all cases the reaction rates
The use of a silver salt to generate the active catalyst provides increased with electron-donating substituted styrenes, i.e. the 4-
an easy way to study the effect of the counter ion. Since L1 gives methoxy- and 4-methylstyrenes, while the electron-withdrawing
the best results and the copper(I) complex is the active form of substituted styrenes, such as 4-chloro- and 3-nitrostyrenes, de-
the catalyst and requires no activation, [Cu(L1)Cl] was used as creased the reaction rate. In chloroform, good r correlations were
the precursor for our further study. The results with different coun- obtained with q = 1.35 for OTf (Fig. 4), 1.36 for both BF4 and
terions and different solvents are listed in Table 5. With triflate as SbF6 and 1.65 for PF6 (Figs. S1–S3). The negative value of q indi-
the anion, trans/cis ratios, enantioselectivies and yields in different cates the formation of an electrophilic metal–carbene complex
solvents of different polarity were similar (entries 1–4). No signif- intermediate. The similarity in q values obtained with catalysts hav-
icant solvent effect was observed. However, when the coordinating ing different counterions, together with the similarity in enantiose-
solvent acetonitrile was used, no reaction occurred (entry 5). lectivities obtained with these catalysts, provide evidence that the
Catalysts with OTf, BF4, PF6 and SbF6 as counterions gave counterion is not involved in the active intermediate.
similar trans/cis ratios, enantioselectivities and isolated yields, no
matter whether the reaction took place in chloroform or THF (en- 3.7. Kinetic rate law
tries 1, 4, 6–11). These results suggest the existence of a common
intermediate for the catalysts with different counterions. For the Kinetic studies for the copper-catalyzed cyclopropanation of
catalyst with ClO4, no reaction took place (entry 12), just like olefins have been reported previously [35–38]. These studies were
 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186 185

-4
4-MeO
0.4 ρ = -1.35
4-Me
-5
H
0.0 4-Cl
log(kx/kH)

ln[EDA]
-6
-0.4
-7
-0.8 3-NO2
-8

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 0 5 10 15 20

Hammett constant σ Time (min)
Fig. 4. Hammett plot for the cyclopropanation of styrene with EDA in CHCl3 using Fig. 7. EDA decomposition in the absence of olefin, catalyzed by [Cu(L)X] (X = OTf
[Cu(L1)OTf] as the catalyst. (d), BF4 (N), PF6 (j) and SbF6 ()).

-4
1 26
-5 kobsd = 0.126 min
24

1/kobsd (min)
-6 22
ln[EDA]

20
-7
18
-8
16
-9 0.04 0.08 0.12 0.16 0.20
0 5 10 15 20 25 30
Time (min) [styrene] (M)

Fig. 5. EDA decomposition catalyzed by [Cu(L1)OTf]. Fig. 8. Variation of kobsd versus different concentrations of styrene for EDA
decomposition using 1 mol% of [Cu(L1)OTf].

H H [Cu(L)X]
0.3
+ X -X
EtO2C CO2Et
+ [Cu(L)]
0.2 EtO2C H
kobsd (min )
-1

EDA
H CO2Et

0.1 k2 k1

EDA N2
H
0.0 0.5 1.0 1.5 2.0 2.5 (L)Cu
mol% of Cu catalyst CO2Et

Fig. 6. Variation of kobsd versus [Cu]total for EDA decomposition using [Cu(L1)OTf]. Scheme 2.

k1 ½Cutotal ½EDA
based on achiral catalysts while we employed chiral catalysts [39]. Rate ¼
1 þ K½styrene
In the absence of styrene, the rates of EDA decomposition were
found to be first order in both [EDA] (Fig. 5) and [Cu]total (Fig. 6), The equilibrium constant K is calculated to be 2.4. The relatively
similar to the previous study by Pérez et al. [37,38]. A comparison small value obtained here when compared with reported value
of the k1 values indicated our catalyst (k1 = 658 M1 min1) is more with Cu(I)Bp (K = 77 ± 29) [38] indicates that [Cu(L)OTf] is mostly
active than the reported Cu(I)Bp catalyst (where Bp = dihydrid- in the catalytically active form during cyclopropanation. This
ibis(1-pyrazoyl)borate) (k1 = 52 M1 min1) [37,38]. With catalysts means more concentrated styrene solutions can be used with our
having different counterions, the rates of EDA decomposition were catalyst. Previously Kochi et al. [35] had also observed similar
also found to be affected strongly by the counterions (Fig. 7). EDA binding and showed catalytically effective copper(I) species were
decomposition rates were the fastest for SbF6 (1.91 min1), deactivated by the multiple coordination of olefin, which in turn
slower for BF4 (1.01 min1) and PF6 (1.18 min1), and slowest inhibited the EDA decomposition.
for OTf (0.126 min1). These results seem to indicate the presence
of an equilibrium between the catalyst and the counterion. With 3.8. Mechanism of copper-catalyzed asymmetric cyclopropanation
these results, a scheme similar to the one proposed by Pérez
et al. is proposed for EDA decomposition (Scheme 2). A general mechanistic scheme with copper-catalyzed cyclo-
In the presence of excess styrene, the reaction rates were found propanation, similar to that proposed by Pérez and his co-workers
to decrease with [styrene]. A plot of 1/kobsd versus [styrene] is [38], is shown in Scheme 3. The active catalyst here, a 14-electron
linear (Fig. 8). The rate law for cyclopropanation under pseudo-first fragment of [Cu(L1)]+, can undergo two possible reactions: (i) with
order conditions is therefore styrene to form a copper–olefin complex, which has a retardation
186 W.-S. Lee et al. / Polyhedron 30 (2011) 178–186

+ These data can be obtained free of charge via http://www.ccdc.ca-

(L1)Cu X (L1)Cu X- m.ac.uk/conts/retrieving.html, or from the Cambridge Crystallo-
Ph
graphic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK;
-X- - styrene K fax: (+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk. Sup-
+X- + styrene plementary data associated with this article can be found, in the
[Cu(L1)]+ online version, at doi:10.1016/j.poly.2010.10.021.

Fumarate
+ References
Maleate Ph COOEt
[1] D.R. McMillin, K.M. McNett, Chem. Rev. 98 (1998) 1201.
EDA k1 [2] R.M. Williams, L. De Cola, F. Hartl, J.-J. Lagref, J.-M. Planeix, A. De Cian, M.W.
Hosseini, Coord. Chem. Rev. 230 (2002) 253.
[3] C. Kaes, A. Katz, M.W. Hosseini, Chem. Rev. 100 (2000) 3553.
EDA [4] R. Ziessel, Coord. Chem. Rev. 216 (2001) 195.
+ Ph
H [5] A. Ion, M. Buda, J.-C. Moutet, E. Saint-Aman, G. Royal, I. Gautier-Luneau, M.
(L1)Cu Bonin, R. Ziessel, Eur. J. Inorg. Chem. (2002) 1357.
CO 2 Et [6] S. Kume, M. Kurihara, H. Nishihara, Inorg. Chem. 42 (2003) 2194.
[7] L.A. Adrio, K.K. Hii, Chem. Commun. (2008) 2325.
[8] J. Niu, H. Zhou, Z. Li, J. Xu, S. Hu, J. Org. Chem. 73 (2008) 7814.
Scheme 3.
[9] C. Ricardo, L.M. Matosziuk, J.D. Evanseck, T. Pintauer, Inorg. Chem. 48 (2009)
16.
effect on EDA decomposition; and (ii) with EDA to form a copper– [10] K. Ito, S. Tabuchi, T. Katsuki, Synlett (1992) 5755.
[11] H.-L. Kwong, W.-S. Lee, H.-F. Ng, W.-H. Chiu, W.-T. Wong, J. Chem. Soc., Dalton
carbene intermediate, which subsequently reacts with styrene to Trans. (1998) 1043.
form cyclopropane or with another equivalent of EDA to form [12] R. Rios, J. Liang, M.M.-C. Lo, G.C. Fu, Chem. Commun. (2000) 377.
diethyl fumarate and maleate as byproducts. [13] D. Lötscher, S. Rupprecht, H. Stoeckli-Evans, A. von Zelewsky, Tetrahedron:
Asymm. 11 (2000) 4341.
[14] A.V. Malkov, M. Bella, V. Langer, P. Kočovský, Org. Lett. 2 (2000) 3047.
4. Conclusions [15] W.-S. Lee, H.-L. Kwong, H.-L. Chan, W.-W. Choi, L.-Y. Ng, Tetrahedron: Asymm.
12 (2001) 1007.
[16] D.R. Boyd, N.D. Sharma, L. Sbircea, D. Murphy, T. Belhocine, J.F. Malone, S.L.
In summary, we have described the synthesis and characteriza- James, C.C.R. Allen, J.T.G. Hamilton, Chem. Commun. (2008) 5535.
tions of a family of copper(II) complexes, [Cu(L)Cl2], derived from [17] C. Bolm, M. Ewald, M. Felder, G. Schlingloff, Chem. Ber. 125 (1992) 1169.
chiral bipyridines L1–6. Our data show the steric properties of [18] H. Fritschi, U. Leutenegger, A. Pfaltz, Helv. Chim. Acta 71 (1988) 1553.
[19] G.G. Mohamed, N.E.A. El-Gamel, Spectrochim. Acta A 61 (2005) 1089.
the ligands not only affect the colour and coordination geometries [20] M. Ghosh, P. Biswas, U. Flörke, K. Nag, Inorg. Chem. 47 (2008) 281.
but also the redox properties. [Cu(L)Cl2] are isolated as either or- [21] A.B.P. Lever, Inorganic Electronic Spectroscopy, second ed., Elsevier,
ange, yellow or green solids. As revealed by X-ray crystallography, Amsterdam, 1984.
[22] G. Murphy, P. Nagle, B. Murphy, B. Hathaway, J. Chem. Soc., Dalton Trans.
[Cu(L5)Cl2] and [Cu(L6)Cl2] have distorted trigonal bipyramidal
(1997) 2645.
geometries that are different from that of [Cu(L1)Cl2]. Cyclic vol- [23] B.J. Hathaway, in: G. Wilkinson, R.D. Gillard, J.A. McCleverty (Eds.),
tammetry data show that complexes with ligands containing bulk- Comprehensive Coordination Chemistry, vol. 5, Pergamon, Oxford, 1987, p.
534.
ier groups (L1–3) are more easily reduced than those with ligands
[24] P. Biswas, S. Dutta, M. Ghosh, Polyhedron 27 (2008) 2105.
containing less bulky group (L4–6). The OTf, PF6, BF4 and SbF6 [25] W.M. Davis, A. Zask, K. Nakanishi, S.J. Lippard, Inorg. Chem. 24 (1985) 3737.
forms of the complex can function as catalysts for the asymmetric [26] R. Mukherjee, in: J.A. McCleverty, T.J. Meyer (Eds.), Comprehensive
cyclopropanation of styrene with EDA. Kinetic and mechanistic Coordination Chemistry II, vol. 6, Elsevier, Amsterdam, 2004, p. 747.
[27] H. Börzel, P. Comba, K.S. Hagen, C. Katsichtis, H. Pritzkow, Chem. Eur. J. 6
studies suggest that a 14-electron species is the active catalyst in (2000) 914.
the carbene transfer reaction. [28] C. O’Sullivan, G. Murphy, B. Murphy, B. Hathaway, J. Chem. Soc., Dalton Trans.
(1999) 1835.
[29] R.D. Willett, G. Pon, C. Nagy, Inorg. Chem. 40 (2001) 4342.
Acknowledgements [30] A.W. Addison, T.N. Rao, J. Reedijk, J. van Rijn, G.C. Verschoor, J. Chem. Soc.,
Dalton Trans. (1984) 1349.
Financial support from the Hong Kong Research Grants Council [31] M. Munakata, M. Maekawa, S. Kitagawa, S. Matsuyama, H. Masuda, Inorg.
Chem. 28 (1989) 4300.
GRF (CityU 101108) and the Area of Excellence Scheme established [32] A.J. Pallenberg, K.S. Koenig, D.M. Barnhart, Inorg. Chem. 34 (1995) 2833.
under the University Grants Committee of the Hong Kong SAR, [33] A.J. Blake, P. Hubberstey, W.-S. Li, D.J. Quinlan, C.E. Russell, C.L. Sampson, J.
China (Project No. AoE/P-10/01) and the City University of Hong Chem. Soc., Dalton Trans. (1999) 4261.
[34] B.A. Gandhi, O. Green, J.N. Burstyn, Inorg. Chem. 48 (2007) 3816.
Kong are gratefully acknowledged.
[35] R.G. Solomon, J.K. Kochi, J. Am. Chem. Soc. 95 (1973) 3300.
[36] M.M. Díaz-Requejo, P.J. Pérez, M. Brookhart, J.L. Templeton, Organometallics
Appendix A. Supplementary data 16 (1997) 4399.
[37] M.M. Díaz-Requejo, M.C. Nicasio, P.J. Pérez, Organometallics 17 (1998) 3051.
[38] M.M. Díaz-Requejo, T.R. Belderrain, M.C. Nicasio, F. Prieto, P.J. Pérez,
CCDC 158090, 732089 and 732090 contains the supplementary Organometallics 18 (1999) 2601.
crystallographic data for [Cu(L1)Cl], [Cu(L5)Cl2] and [Cu(L6)Cl2]. [39] For the derivation of the rate law, see supporting information.