Catalysis of alkene epoxidation by a series of gallium(III) complexes with neutral N-donor ligands

Article abstract of DOI:10.1021/ic400570h

Six gallium(III) complexes with N-donor ligands were synthesized to study the mechanism of Ga^III-catalyzed olefin epoxidation. These include 2:1 ligand/metal complexes with the bidentate ligands ethylenediamine, 5-nitro-1,10-phenanthroline, and 5-amino-1,10-phenanthroline, as well as 1:1 ligand/metal complexes with the tetradentate N,N′-bis(2-pyridylmethyl)-1, 2-ethanediamine, the potentially pentadentate N,N,N′-tris(2-pyridylmethyl) -1,2-ethanediamine, and the potentially hexadentate N,N,N′,N′- tetrakis(2-pyridylmethyl)-1,2-ethanediamine. In solution, each of the three pyridylamine ligands appears to coordinate to the Ga^III through four donor atoms. The six complexes were tested for their ability to catalyze the epoxidation of alkenes by peracetic acid. Although the complexes with relatively electron-poor phenanthroline derivatives display faster initial reactivity, the gallium(III) complexes with the polydentate pyridylamine ligands appear to be more robust, with less noticeable decreases in their catalytic activity over time. The more highly chelating trispicen and tpen are associated with markedly decreased activity.

Full text of DOI:10.1021/ic400570h

Article
pubs.acs.org/IC
Catalysis of Alkene Epoxidation by a Series of Gallium(III) Complexes
with Neutral N‑Donor Ligands
Wenchan Jiang, John D. Gorden, and Christian R. Goldsmith*
Department of Chemistry and Biochemistry, Auburn University, Auburn, Alabama 36849, United States
*
S Supporting Information
ABSTRACT: Six gallium(III) complexes with N-donor ligands were
III
synthesized to study the mechanism of Ga -catalyzed olefin epoxidation.
These include 2:1 ligand/metal complexes with the bidentate ligands
ethylenediamine, 5-nitro-1,10-phenanthroline, and 5-amino-1,10-phenanthro-
line, as well as 1:1 ligand/metal complexes with the tetradentate N,N′-bis(2-
pyridylmethyl)-1,2-ethanediamine, the potentially pentadentate N,N,N′-
tris(2-pyridylmethyl)-1,2-ethanediamine, and the potentially hexadentate
N,N,N′,N′-tetrakis(2-pyridylmethyl)-1,2-ethanediamine. In solution, each of
III
the three pyridylamine ligands appears to coordinate to the Ga through four
donor atoms. The six complexes were tested for their ability to catalyze the
epoxidation of alkenes by peracetic acid. Although the complexes with
relatively electron-poor phenanthroline derivatives display faster initial
reactivity, the gallium(III) complexes with the polydentate pyridylamine
ligands appear to be more robust, with less noticeable decreases in their
catalytic activity over time. The more highly chelating trispicen and tpen are associated with markedly decreased activity.
INTRODUCTION
group 13 metals, this chemistry has never been systematically
studied.
Here, we report the syntheses and characterizations of six
■
The coordination chemistry of gallium has been investigated to
serve a number of applications, including the development of
volatile precursors for gallium oxide, nitride, selenide, arsenide,
gallium(III) complexes with neutral N-donor ligands (Scheme
1
). The ligands were selected to interrogate the influences of
1
−7
and sulfide layers in chemical vapor deposition
as well as
ligand denticity and electronics on the catalytic activity. One
6
8
Ga-containing radiopharmaceuticals suitable for positron
+
noted shortcoming with the [Ga(phen) Cl ] system is that its
2
2
8
−11
20
emission tomography.
Gallium(III) compounds have also
catalysis essentially ends after approximately 1 h. Replacing
the two bidentate phen ligands with a single more highly
coordinating molecule extends the lifetime of the catalysis and
increases the optimum yield of the epoxides, although the
activity decreases if the ligand has the capacity to bind in a
penta- or hexadentate fashion. The more electron-poor ligands
promote more extensive alkene epoxidation, with the best
activity being associated with 5-nitro-1,10-phenanthroline
garnered attention for their ability to act as homogeneous Lewis
7
,12−16
acid catalysts.
One trend observed in the aforementioned
catalysis is that cationic gallium(III) species tend to promote
faster and more extensive reactivity.
Recently, our group reported that the previously known
compound [Ga(phen) Cl ]Cl (phen = 1,10-phenanthro-
2
2
1
7−19
line)
could catalyze the epoxidation of alkenes by peracetic
2
0
(
NO -phen).
acid. This is a rare example of gallium accelerating an
2
21−24
oxidation−reduction reaction
and, to the best of our
III
knowledge, represents the first instance of a homogeneous Ga
catalyst for olefin epoxidation. The reactivity proceeded more
quickly and at a lower temperature than analogous reactions
EXPERIMENTAL SECTION
■
Materials. Cyclooctene, cyclohexene, styrene, 1-octene, 4-vinyl-
cyclohexene, 1,2-ethanediamine (en), 5-amino-1,10-phenanthroline
2
5−27
catalyzed by aluminum(III) complexes.
We attributed the
(NH
-phen), 5-nitro-1,10-phenanthroline (NO -phen), and gallium
2 2
+
trichloride (GaCl ) were purchased from Sigma-Aldrich and used as
activity of [Ga(phen) Cl ] to the presence of the neutral and
3
2
2
28−30
received. Methylene chloride (CH Cl ) was bought from Macron
2 2
relatively electron-poor phen ligands,
remain bound to the Ga during the catalysis. Without these
ligands, oxidation does not proceed; GaCl by itself is not a
which appear to
Chemicals. Anhydrous acetonitrile (MeCN) and methanol (MeOH)
were procured from Acros. Diethyl ether (ether) was obtained from J.
III
20
3
T. Baker. Deuterated chloroform (CDCl
), acetonitrile (CD CN),
3 3
competent catalyst. Neutral N-donor ligands do not commonly
dimethyl sulfoxide (DMSO-d ), water (D O), and methanol
6
2
III
(
CD OD) were purchased from Cambridge Isotopes.
coordinate to Al , which instead prefers to ligate harder bases.
3
III
These ligands’ affinities for Ga , however, are considerably
3
1
III
higher. Although complexation of Ga to polydentate neutral
N-donor ligands could result in enhanced catalytic activity for
Received: December 13, 2012
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1
Peracetic acid (PAA ) was prepared through a previously described
R
Table 1. Select Crystallographic Data for [Ga(bispicen)Cl2]
procedure that uses an acidic resin as the catalyst in place of the
Cl (1) and [Ga(trispicen)Cl](GaCl )(Cl) (7)
32
4
commercially used sulfuric acid. 50% H O (17 g, 0.25 mol) was
2
2
slowly added to 150 g of glacial acetic acid (2.5 mol) in a room
temperature (RT) polyethylene bottle. After the addition was
complete, 5.0 g of Amberlite IR-120 was added, and the reaction
mixture was stirred behind a blast shield for 24 h at RT. After that
time, the solution was filtered to remove the resin and yield the 7.5 wt
[Ga(bispicen)Cl
2
]
[Ga(trispicen)Cl](GaCl )
4
parameter
formula
Cl·3H O
(Cl)·MeCN
2
C H Cl GaN O
Ga H Cl Ga N
6
14
24
3
4
3
22 26
6
2
MW
472.45
monoclinic
727.64
triclinic
cryst syst
space group
a (Å)
%
PAA solution used for the reactivity experiments. The solution was
R
P2 /c(No. 14)
1
̅
P1 (No. 2)
stored in a freezer when not in use. The concentration of PAA was
R
13.7249(5)
7.3885(3)
20.7913(8)
90
9.2381(7)
12.0453(9)
14.7889(11)
88.9220(10)
79.7160(10)
69.0670(10)
1510.5(2)
2
determined by comparing the intensities of the 13C NMR resonances
of CH CO H and CH CO H.
b (Å)
3
3
3
2
c (Å)
Caution! Peracids and their mixtures with organic solvents are
potentially explosive and should be handled judiciously. The potential
hazards can be minimized by using minimal amounts of these materials at
lower temperatures and by using proper protective equipment, such as blast
shields.
α (deg)
β (deg)
γ (deg)
V (Å3)
105.600(1)
90
2030.70(14)
4
1
Instrumentation. All H NMR spectra were acquired on either a
Z
2
50, 400, or 600 MHz AV Bruker NMR spectrometer at 294 K. All
resonances were referenced to internal standards. Atlantic Microlabs
Norcross, GA) performed all elemental analyses. A Shimadzu IR
cryst color
T (K)
colorless
296
colorless
296
(
reflns collected
unique reflns
63556
59468
Prestige-21 FT-IR spectrophotometer was used for the described IR
spectroscopy. Gas chromatography (GC) data were obtained using a
ThermoScientific Trace GC Ultra spectrometer with a flame ionization
detector. High-resolution mass spectrometry (HR-MS) data were
collected at the Mass Spectrometer Center at Auburn University using
a Bruker microflex LT MALDI-TOF mass spectrometer via direct
probe analysis operated in the positive-ion mode.
4194
7113
R1(F) [I >
0.0301
0.0293
a
2
σ(I)]
2
wR2(F ) (all
0.0828
0.0708
a
data)
a
R1 = ∑||F | − |F ||/∑|F |. wR2 = [∑w(F − F ) /∑w(F ) ]1/2.
2
o
2
c
2
2 2
o
o
c
o
X-ray Crystallography. X-ray diffraction data were acquired using
a Bruker SMART APEX CCD X-ray diffractometer and Mo Kα
radiation. Crystalline samples were mounted in Paratone-N oil on glass
fibers. The program SMART (version 5.624) was used for the
preliminary determination of cell constants and data collection control.
Determination of the integrated intensities and global cell refinement
was performed with the Bruker SAINT software package using a
narrow-frame integration algorithm. The program suite SHELXTL
at RT under N . After 2 h, ether was added to the cloudy solution,
precipitating 0.360 g of the product as a white powder (82%).
2
Colorless crystals were grown from the slow diffusion of ether into a
1
saturated solution of the complex in MeOH. H NMR (CD OD, 400
3
MHz): δ 9.41 (2H, d, J = 5.2 Hz), 8.24 (2H, t, J = 7.6 Hz), 7.77 (2H, t,
J = 6.4 Hz), 7.71 (2H, d, J = 7.6 Hz), 4.88 (2H, d, J = 17.2 Hz), 4.18
(
2H, d, J = 17.2 Hz), 2.69 (2H, d, J = 9.2 Hz), 2.53 (2H, d, J = 9.2
13
(
version 5.1) was used for space group determination, structure
Hz). C NMR (CD OD, 150 MHz): δ 153.7, 147.0, 143.0, 125.8,
1
3
33
2
−1
solution, and refinement. Refinement was performed against F by
24.6, 51.5, 47.0. IR (KBr, cm ): 3416 (s), 3389 (s), 3100 (s), 3040
weighted full-matrix least squares, and empirical absorption corrections
(s), 2991 (s), 2914 (s), 2850 (s), 2482 (w), 2395 (w), 2270 (w), 2166
(w), 1665 (w), 1612 (s), 1578 (m), 1481 (s), 1440 (s), 1363 (s), 1300
(s) 1262 (m), 1226 (s), 1160 (m), 1099 (s), 1048 (s), 1033 (s), 980
(m), 906 (w), 849 (m), 821 (w), 777 (s), 723 (m), 649 (m), 567 (m),
506 (m), 480 (m), 425 (m). Elem anal. Calcd for
C H N GaCl ·H O: C, 38.53; H, 4.62; N, 12.84. Found: C, 38.32;
34
(
SADABS) were applied. H atoms were placed at calculated
positions using suitable riding models with isotropic displacement
parameters derived from their carrier atoms. Crystallographic data are
provided in both Table 1 and the Supporting Information.
Syntheses. The ligands N,N′-bis(2-pyridylmethyl)-1,2-ethanedi-
14
18
4
3
2
amine (bispicen), N,N,N′-tris(2-pyridylmethyl)-1,2-ethanediamine
H, 4.53; N, 12.51.
Dichloro[N,N,N′-tris(2-pyridylmethyl)-1,2-ethanediamine]-
gallium(III) Chloride (2). GaCl (0.189 g, 1.074 mmol) and trispicen
(0.359 g, 1.079 mmol) were dissolved in 10 mL of CH Cl (2 mL).
The resultant solution was stirred for 2 h at RT under N . The
(
trispicen), and N,N,N′,N′-tetrakis(2-pyridylmethyl)-1,2-ethanedi-
35−37
amine (tpen) were prepared as described previously.
The
3
identities and purities of these compounds were confirmed by NMR.
cis-Dichloro[N,N′-bis(2-pyridylmethyl)-1,2-ethanediamine]-
2
2
2
gallium(III) Chloride (1). GaCl (0.164 g, 0.931 mmol) and bispicen
addition of ether deposited the product as a pale-yellow powder (0.356
3
1
(
0.227 g, 0.937 mmol) were dissolved in 10 mL of CH Cl and stirred
g, 65%). H NMR (CD OD, 400 MHz): δ 9.40 (1H, d, J = 4.4 Hz),
2
2
3
B
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
9
(
.17 (1H, d, J = 4.4 Hz), 8.45 (1H, d, J = 5.2 Hz), 8.26 (2H, m), 7.96
2H, m), 7.83 (2H, m), 7.61 (1H, d, J = 5.2 Hz), 7.55 (1H, t, J = 4.4
Hz), 7.16 (1H, d, J = 3.6 Hz), 4.56 (1H, d, J = 15.6 Hz), 4.49 (1H, d, J
d, J = 7.8 Hz), 8.75 (1H, m), 8.46 (1H, m), 8.27 (1H, d, J = 7.8 Hz),
8.21 (1H, m), 7.66 (2H, m), 7.40 (1H, m), 7.25 (2H, m), 7.14 (1H, s).
13
C NMR (DMSO-d , 62.5 MHz): δ 148.2, 145.1, 144.9, 144.7, 144.5,
6
=
15.6 Hz), 4.08 (1H, d, J = 14.4 Hz), 3.58 (4H, m), 3.20 (1H, m),
142.6, 138.9, 138.3, 137.2, 136.9, 136.4, 136.0, 132.4, 132.0, 129.6,
128.8, 126.6, 126.3, 125.5, 125.2, 122.8, 122.3, 101.0, 100.8. IR (KBr,
2
1
1
.93 (1H, m), 2.84 (1H, m). 13C NMR (CD OD, 150 MHz): δ 148.5,
3
−
1
47.9, 147.5, 146.0, 144.7, 144.6, 143.5, 128.8, 128.1, 127.9, 127.1,
26.3, 126.2, 124.7, 61.2, 61.1, 56.8, 53.1, 47.5. IR (KBr, cm ): 3487
cm ): 3399 (s), 3339 (s), 3208 (s), 3080 (m), 1640 (s), 1620 (s),
1604 (s), 1586 (m), 1522 (w), 1494 (s), 1465 (m), 1435 (s), 1350
(w), 1323 (w), 1287 (w), 1226 (w), 1164 (w), 1146 (w), 1122 (w),
1087 (w), 913 (w), 851 (w), 826 (w), 802 (w), 725 (s), 665 (w), 653
(w), 427 (w). Elem anal. Calcd for C H N GaCl ·H O: C, 49.31; H,
−1
(
(
s), 3387 (s), 3334 (s), 3115 (s), 2850 (s), 2657 (s), 2476 (m), 2392
m), 2271 (m), 2157 (m), 2034 (w), 1936 (w), 1882 (w), 1815 (w),
1
1
747 (w), 1611 (s), 1478 (s), 1438 (s), 1363 (s), 1298 (s), 1230 (s),
159 (s), 1107 (s), 1040 (s), 1031 (s), 977 (s), 910 (w), 849 (m), 777
24
18
6
3
2
3.45; N, 14.38. Found: C, 49.56; H, 3.43; N, 14.51.
(
s), 720 (m), 646 (m), 565 (m), 492 (m), 424 (m). Elem anal. Calcd
Dichlorobis(1,2-ethanediamine)gallium(III) Chloride (6). GaCl3
(0.214 g, 1.22 mmol) and ethylenediamine (165 μL, 2.46 mmol)
for C H N GaCl : C, 47.15; H, 4.55; N, 13.75. Found: C, 47.44; H,
20
23
5
3
4.45; N, 13.36.
were dissolved in 10 mL of MeCN under N . As the reaction was
2
Dichloro[N,N,N′,N′-tetrakis(2-pyridylmethyl)-1,2-
stirred for 2 h, a white precipitate began to form. The addition of ether
ethanediamine]gallium(III) Chloride (3). GaCl3 (0.134 g, 0.761
mmol) and tpen (0.324 g, 0.763 mmol) were dissolved in 10 mL of
CH Cl under N . The resultant solution was stirred for 2 h at RT,
deposited more solid. A total of 0.320 g of the product was isolated as
1
a white powder (79%). H NMR (DMSO-d , 400 MHz): δ 3.07 (8H,
6
1
13
2
2
2
s). H NMR (D O, 600 MHz): δ 3.14 (8H, s). C NMR (DMSO-d6,
100 MHz): δ 36.57. C NMR (D O, 150 MHz): δ 39.31. IR (KBr,
cm ): 3420 (s), 3225 (s), 2978 (s), 2912 (s), 2804 (s), 2747 (s),
2
13
after which excess ether was added to precipitate 0.388 g of the
2
1
−1
product as a white powder (85%). H NMR (CDCl , 250 MHz): δ
3
9
.61 (2H, d, J = 5.4 Hz), 8.64 (2H, d, J = 4.8 Hz), 8.19 (2H, t, J = 8.2
2711 (s), 2677 (s), 2634 (s), 2575 (m), 2521 (m), 2420 (m), 2280
(w), 2055 (m), 1624 (m), 1602 (m), 1504 (s), 1342 (w), 1182 (w),
1085 (m), 1034 (s), 1008 (w), 973 (w), 917 (w), 786 (w), 568 (m),
467 (m), 445 (w), 431 (w). Elem anal. Calcd for
C H N GaCl ·2H O: C, 14.46; H, 6.07; N, 16.86. Found: C, 14.61;
Hz), 7.85 (2H, m), 7.74 (2H, t, J = 6.8 Hz), 7.60 (4H, m), 7.39 (2H,
m), 5.09 (2H, d, J = 16.2 Hz), 4.65 (2H, d, J = 13.2 Hz), 4.20 (2H, d, J
=
MHz): δ 9.52 (2H, d, J = 5.2 Hz), 8.63 (2H, d, J = 5.6 Hz), 8.21 (2H,
t, J = 8.0 Hz), 7.83 (2H, t, J = 5.6 Hz), 7.75 (2H, t, J = 6.4 Hz), 7.62
1
11.4 Hz), 3.60 (4H, m), 3.00 (2H, m). H NMR (CD CN, 400
3
8
16
4
3
2
H, 6.03; N, 16.71.
Reactivity Studies. All reactions were run in MeCN at 273 K
under N , with the initial concentrations of the catalyst, alkene, and
PAA being 5.0 mM, 500 mM, and 1.0 M, respectively. Most reactions
(
4H, d, J = 7.6 Hz), 7.41 (2H, m), 4.93 (2H, d, J = 16.0 Hz), 4.56 (2H,
d, J = 13.6 Hz), 3.97 (2H, d, J = 16.4 Hz), 3.61 (2H, d, J = 13.6 Hz),
2
3
1
1
2
2
.48 (2H, d, J = 5.2 Hz), 3.27 (2H, d, J = 5.2 Hz). 13C NMR (CDCl3,
00 MHz): δ 151.9, 151.5, 149.9, 147.0, 142.8, 137.5, 127.5, 126.6,
25.6, 124.2, 56.6, 56.4, 46.3. IR (KBr, cm ): 3045 (m), 3008 (s),
945 (s), 2907 (s), 2863 (s), 2815 (s), 2800 (s), 2764 (s), 2695 (m),
550 (w), 2300 (w), 1905 (w), 1859 (w), 1796 (w), 1662 (w), 1588
R
were allowed to proceed for 1 h. At the end of each reaction, excess
ether was added to precipitate the gallium(III) compound, effectively
quenching catalysis. For the studies focusing on the catalytic activity as
a function of time, aliquots were taken at the indicated time points.
Each portion was quenched with excess ether and filtered through a
−1
(
(
(
s), 1572 (s), 1472 (s), 1433 (s), 1364 (s), 1345 (s), 1298 (s), 1245
s), 1209 (m), 1170 (s), 1126 (s), 1085 (m), 1045 (m), 1016 (w), 989
s), 902 (m), 864 (m), 839 (m), 786 (s), 765 (s), 732 (m), 687 (w),
plug of silica gel to remove the remaining PAA and gallium(III) salts.
R
Control studies confirmed that this workup removed neither the olefin
starting material nor the epoxide product from the solution. The
remaining alkene and epoxide products were quantified relative to an
internal standard of 1,2-dichlorobenzene, which is inert under the
reaction conditions. All reactions were run at least three times to
ensure reproducibility. All reported values are the averages of the
results from these reactions. The errors represent 1 standard deviation.
662 (w), 617 (m), 593 (w), 546 (w), 470 (w). MS (ESI). Calcd for
+
[
Ga(tpen)Cl ] : m/z 563.1008. Found: m/z 563.0988. Elem anal.
2
Calcd for C H N GaCl : C, 51.99; H, 4.70; N, 13.99. Found: C,
26
28
6
3
5
1.70; H, 4.65; N, 14.00.
Dichlorobis[5-nitro(1,10-phenanthroline)]gallium(III) Chloride
4). GaCl3 (0.169 g, 0.959 mmol) and NO -phen (0.437 g, 1.94
(
2
mmol) were combined in 10 mL of CH Cl under N . After the
2
2
2
solution was stirred for 2 h at RT, ether was added, precipitating 0.332
RESULTS
1
■
g of the product as a pale-yellow powder (51%). H NMR (CD OD,
3
Synthesis and Characterization. Syntheses of the
gallium(III) complexes (Scheme 1) proceed in a straightfor-
ward manner, and simple mixing of the GaCl
6
=
(
00 MHz): δ 10.35 (2H, m), 9.82 (1H, d, J = 12.0 Hz), 9.50 (1H, d, J
8.4 Hz), 9.44 (1H, s), 9.33 (2H, m), 9.03 (1H, d, J = 8.4 Hz), 8.67
2H, m), 8.06 (1H, d, J = 3.6 Hz), 8.02 (1H, d, J = 3.6 Hz), 7.86 (2H,
salt and ligand in
3
13
m). C NMR (CD OD, 150 MHz): δ 155.9, 153.4, 151.8, 149.6,
RT CH Cl or MeCN is sufficient to ensure complexation. The
3
2
2
1
1
1
2
48.0, 147.2, 146.6, 146.3, 146.0, 144.6, 141.4, 140.0, 139.5, 139.0,
compounds with the polydentate pyridylamine ligands
bispicen, trispicen, and tpen) can be isolated in high purities
33.2, 131.1, 129.6, 129.3, 129.0, 128.4, 128.1, 127.1, 126.8, 124.9,
(
−1
24.5. IR (KBr, cm ): 3422 (m), 3113 (w), 3071 (m), 3004 (w),
970 (w), 2934 (w), 2871 (w), 2337 (w), 1970 (w), 1946 (w), 1841
and yields (77−85%) through precipitation from CH Cl /ether
2
2
mixtures. Compounds 1−3 tend to be hygroscopic and
noticeably moistened upon prolonged exposure to air. The
compounds with the bidentate ligands (NO -phen, NH -phen,
(
1
(
9
w), 1627 (m), 1586 (m), 1540 (s), 1522 (s), 1507 (s), 1490 (m),
453 (m), 1421 (s), 1388 (m), 1349 (s), 1335 (s), 1261 (w), 1207
m), 1181 (m), 1153 (m), 1117 (m), 1104 (m), 1038 (w), 992 (w),
74 (w), 920 (m), 838 (s), 823 (s), 810 (m), 751 (m), 734 (s), 721
s), 652 (m), 619 (w), 541 (w), 507 (w), 429 (w). Elem anal. Calcd
2
2
and en) are prepared in moderate-to-high yields (51−87%)
through precipitation. The en complex 6 is the least soluble of
the six gallium(III) compounds in organic solvents. Complexes
4−6 appear to be more hygroscopic than compounds 1−3, and
we were unable to exclude water from the elemental analyses of
the former. When syntheses of 1−6 were run under air instead
(
for C H N GaCl ·3H O: C, 42.34; H, 2.96; N, 12.35. Found: C,
24
14
6
3
2
42.56; H, 2.73; N, 12.11.
Dichlorobis[5-amino(1,10-phenanthroline)]gallium(III) Chloride
(
5). GaCl3 (0.104 g, 0.589 mmol) and NH -phen (0.228 g, 1.18
2
mmol) were dissolved in 10 mL of MeCN under N . The solution was
2
of N , much lower yields of the desired products were obtained.
2
allowed to stir for 2 h at RT. After that time, excess ether was added to
The gallium(III) complexes are diamagnetic and nearly
colorless, as anticipated. The 294 K H NMR spectrum of the
bispicen complex 1 has four resonances in the aromatic region,
indicating that its two pyridine rings are equivalent at this
temperature. The doublet at 9.41 ppm is assigned to the
protons on the 6-positions of the pyridine rings. These
1
precipitate 0.287 g of the product as a yellow powder (87%). H NMR
1
(
DMSO-d , 250 MHz): δ 10.00 (1H, t, J = 5.2 Hz), 9.54 (2H, m), 9.06
6
(
1H, d, J = 8.0 Hz), 8.88 (1H, d, J = 8.5 Hz), 8.55 (1H, m), 8.40 (1H,
d, J = 8.2 Hz), 8.28 (1H, m), 7.71 (2H, m), 7.45 (1H, m), 7.25 (2H,
1
m), 7.10 (1H, s). H NMR (CD OD, 600 MHz): δ 10.18 (1H, d, J =
3
16.5 Hz), 9.74 (1H, d, J = 16.2 Hz), 9.39 (1H, t, J = 6.9 Hz), 8.91 (1H,
C
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
resonances tend to be exquisitely sensitive to metal-ion
coordination, and their chemical shifts can be used to assess
whether the affiliated pyridine rings bind to Ga . The
Scheme 3. Proposed Solution Structure of 3
III 38
resonances at 4.88 and 4.18 ppm indicate that the picolylic
methylene protons are diastereotopic. The splitting pattern is
inconsistent with a trans ligand conformation, which has a
sufficiently high symmetry to render all of the picolylic protons
equivalent. The splitting pattern is more consistent with a C2
symmetry for the gallium(III) complex in solution. The 13
C
exchange between the pendant and bound pyridine rings
NMR spectrum features only seven resonances, providing
further evidence that the halves of the bispicen ligand are
equivalent in solution. The NMR data suggest that 1 retains the
cis-α ligand conformation observed in its crystal structure (vida
infra) in solution.
appears to be slow; we observe neither broadening nor
1
coalescing of the H NMR signals as samples of 3 in CD CN
3
are heated from 294 to 324 K.
The solution structures of the two gallium(III) complexes
with the phen derivatives are more difficult to interpret.
1
The H NMR spectrum of the trispicen complex 2 has two
Potentially, two 5-derivatized phen ligands could coordinate to
doublets at 9.40 and 9.17 ppm, which are comparable in energy
to the 9.41 ppm resonance observed for 1. We have therefore
assigned these as corresponding to the protons on the 6-
+
the metal ion to form six isomers of cis-[Ga(R-phen) Cl ] ;
2
2
these six are comprised of three pairs of enantiomers (Scheme
4). Given that the cis conformation is present in the structures
III
positions of Ga -bound pyridine rings. The integrated intensity
of each of these doublets corresponds to one proton, suggesting
that the two associated pyridine rings are inequivalent at 294 K.
The resonance for the third pyridine ring’s 6-position proton
appears at a significantly lower chemical shift (8.45 ppm),
Scheme 4. Possible Conformations of the phen Ligands in
Complexes 4 and 5
III
suggesting that it does not coordinate to Ga . The
inequivalence of the two coordinated pyridine rings is
1
3
supported by the acquired C NMR spectrum. If the pyridine
1
3
rings were equivalent, we would observe 14 C resonances at
most; instead, 19 resonances can be unambiguously identified.
The conformation of the four coordinated N donors cannot be
assigned with certainty; Scheme 2 shows three possible solution
structures in which the two chlorides are coordinated cis to
each other.
Scheme 2. Possible Solution Structures of 2
+
+
20,39
of both [Ga(phen) Cl ] and [Ga(phen) Br ] ,
we suspect
2
2
2
2
The tpen complex 3 likewise appears to have the
that additional trans isomers are not present in significant
pyridylamine ligand coordinated to the metal in a hypodentate
quantities, although the data certainly cannot preclude such a
1
1
13
fashion. The H NMR spectra in CD CN and CDCl strongly
possibility. The H and C NMR spectra of the two
3
3
resemble each other, suggesting that MeCN does not displace
the chlorides to an observable degree. The NMR spectra are
consistent with a structure with two pendant pyridine rings.
This is again most apparent from consideration of the protons
on the 6-positions of the pyridine moieties, which are evenly
split between doublet resonances at ∼9.6 (bound) and ∼8.6
gallium(III) complexes with NO -phen and NH -phen, 4 and
2
2
5, display more resonance features than would be anticipated
from the uncoordinated ligands. This demonstrates that the
isomers of the gallium(III) complexes, like complex 3, have
limited fluxionality at RT. Rapid isomerization of 4 and 5 would
1
result in 7 and 8 observable H NMR resonances, respectively;
(
unbound) ppm. There are four doublets that can be
instead, we observe 10 and 12 features. As with 3, we do not
1
reasonably assigned to methylene protons; as with 1, this
pattern is consistent with a C2 symmetry and a cis-α
conformation of the N donors. The number of resonances in
the 13C NMR data suggests that the halves of the tpen ligand
are equivalent in solution, supporting the assignment of a C2-
symmetric species. HR-MS analysis of 3 detects a [Ga(tpen)-
observe any H NMR resonances broadening or shifting as
samples of 4 are heated to 324 K. Given the number of
1
multiplets observed in the H NMR spectra, we currently
believe that 4 and 5 each exist as a mixture of multiple cis
isomers in solution.
The H and 13C NMR spectra of the en complex 6 each
contain a single resonance, consistent with the presence of a
single compound with two chemically equivalent en ligands.
These NMR spectra are distinct from those of free en in both
1
+
III
Cl ] cation; consequently, we believe that the Ga center in 3,
2
like those in 1 and 2, is primarily coordinated by a N Cl set of
4
2
donor atoms in solution. The mode of tpen coordination that is
most consistent with the data is depicted in Scheme 3. The
DMSO-d6 and D O, indicating that the ligand is indeed
2
D
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
chelating the metal ion. The data in D O suggest that the
2
III
Ga −N bonds remain intact in solvents that can potentially
III
ligate the Ga center.
IR Spectroscopy. IR measurements provide support for the
modes of coordination suggested by the NMR data. The IR
spectrum of noncoordinated pyridine features a C−N stretch at
−
1
1
578 cm ; coordination to a metal ion typically shifts this
−1
40
frequency to the 1590−1615 cm range. All three complexes
have intense bands where the C−N stretches of coordinated
pyridine rings are expected (1612, 1611, and 1588 cm− for 1−
1
3
, respectively). The IR spectra of complexes 1−3 display
−1
stronger absorbances in the 1560−1580 cm region as the
number of pyridine rings increases, with complexes 1 and 3
−1
having weak and strong bands at 1578 and 1572 cm ,
respectively. The presence of this peak for 1 is unexpected and
may result from either another vibrational mode or the partial
dissociation of the ligand during the preparation of the KBr
pellet. Complex 2 lacks a distinct band in the 1560−1580 cm
range, but its 1611 cm feature has a prominent shoulder that
extends into this region.
2
+
Figure 2. ORTEP representation of the dication [Ga(trispicen)Cl] .
−
−
All H atoms, the Cl and [GaCl4] counteranions, and a non-
coordinated MeCN molecule have been removed for clarity. All
thermal ellipsoids are drawn at 50% probability.
−1
−1
freshly isolated solid samples of 2, we believe that 2 at least
partially converts to 7 over the prolonged period of time
required for crystal growth. Only small yields of 7 were
obtained, limiting our ability to characterize the compound and
its reactivity. The IR spectrum of 7 is distinct from that of 2 and
notably lacks a feature between 1560 and 1580 cm− that can
be readily assigned to an uncoordinated pyridine ring. In the
X-ray Crystallography. Crystals of 1 were grown from the
slow diffusion of ether into saturated MeOH solutions (Table
). The X-ray diffraction data confirm that the composition of 1
1
is [Ga(bispicen)Cl ]Cl (Figure 1). The outer-sphere chloride is
2
1
structure of 7, the bound trispicen is κ-5 rather than κ-4. Two
−
different counteranions are in the outer sphere: [GaCl ] and
4
−
Cl . The presence of the tetrachlorogallate(III) anion indicates
III
that some of the Ga in the sample had dissociated from the
trispicen ligand over the time required for crystallization.
The crystals that grew from saturated solutions of 3 were
likewise consistent with decomposition products rather than
the formula anticipated from the NMR and elemental analysis
III
of the freshly precipitated product. In some cases, the Ga
centers in the crystals were coordinated by bispicen, with unit
cell parameters that were identical to those of the
aforementioned crystals of 1. In other cases, gallium(III)
complexes with trispicen were isolated from the solutions. Both
degradation products contain ligands that are missing picolyl
arms. Water molecules are present in these crystal structures,
which may suggest that these portions are being lost through
hydrolytic processes. Attempts to grow the crystals in the strict
absence of air and moisture were unsuccessful.
Figure 1. ORTEP representation of the complex [Ga(bispicen)Cl ]Cl.
2
All H atoms, except for the one located on N(8), and three
noncoordinated water molecules have been removed for clarity. All
thermal ellipsoids are drawn at 50% probability.
́
3
.19 Å from one of the amine N atoms, N(8); this may be
Short-Term Reactivity. The six gallium(III) compounds
were tested for their ability to catalyze the epoxidation of
41
consistent with a N−H···Cl hydrogen bond. In the cation, the
bispicen ligand is bound to Ga in a cis-α conformation, which
III
alkenes by PAA (Table 2). A 1% loading of catalyst and 2
R
is commonly seen in the ligand’s complexes with transition-
equiv of PAA per 1 equiv of substrate were used, primarily to
R
4
2−46
metal ions,
and contributes to an overall distorted
facilitate comparison to our previous results with [Ga-
20
octahedral coordination geometry around the metal center. In
gallium chemistry, the cis-α conformation of four N donors has
also been observed in complexes with N O coordination
(phen) Cl ]Cl. For each catalyzed oxidation reaction, the
2 2
epoxide is the sole organic product observed at 1 h; no allylic
oxidation or dihydroxylation is observed with 1% mole catalytic
loading. In all cases, the percentage of the alkene substrate that
has been consumed is equal within error to the yield of the
epoxide product, suggesting that the GC analysis has accounted
for all organic products. As anticipated, the more electron-
deficient terminal alkenes react less readily, as evidenced by the
lower yields of epoxides.
Three trends can be observed in the epoxidation activities.
First, the gallium(III) complexes with phen derivatives promote
alkene epoxidation to much greater extents. In particular,
compounds 4 and 5 promote the oxidation of 1-octene to 1-
octene oxide to much higher degrees than any of the other four
4
2
1
1,31
spheres.
1
The Ga−N and Ga−Cl bond lengths observed for
are typical for a hexacoordinate gallium(III) complex with a
The Ga−N bond lengths differ
only slightly, and the symmetry of the cation is approximately
2
0,47,48
N Cl donor atom set.
4
2
C .
2
A small number of crystals were also grown from saturated
solutions of 2 in mixtures of MeOH and MeCN (Table 1). The
structural data, however, correspond to the formula [Ga-
(
trispicen)Cl](GaCl )(Cl) (7; Figure 2), as opposed to
4
[
Ga(trispicen)Cl ]Cl. Given that only the latter composition
2
is consistent with the NMR, IR, and elemental analysis data for
E
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Table 2. Alkene Conversions/Yields of Epoxides (%) of Reactions Catalyzed by 1−6
a
III
Standard reaction conditions: MeCN, 273 K, N , [Ga ] = 5.0 mM, [alkene] = 500 mM, and [PAA ] = 1.0 M. The alkene conversion, defined as
2
0
R 0
the percentage of alkene that has been consumed by the reaction, was measured at 1 h via GC. The alkene conversions are italicized in the above
table. The yield of epoxide, defined as the percentage of alkene converted to the epoxide, was measured at 1 h via GC. Because the substrate is in a
b
1
00-fold excess relative to the catalyst, the above yields of epoxides also serve as turnover numbers. From ref 20.
newly reported gallium(III) complexes. The increases in the
yields of the other epoxide products are much less pronounced.
Second, the use of more electron-deficient ligands tends to
result in more epoxidation. Comparison of the activities of
[
Ga(phen) Cl ]Cl, 4, and 5 reveals that the installation of a
2 2
more electron-withdrawing NO group on the 5-positions of
2
the phen ligands increases the turnover rate, whereas the
addition of an electron-donating NH group reduces the yield
2
of the epoxide at 1 h. Third, as the maximum denticity of the
ligand increases past κ-4, the gallium(III) complexes become
less able to quickly catalyze olefin epoxidation. Within the series
of complexes with pyridylamine ligands, the catalytic activity
associated with the tetradentate ligand bispicen is fastest;
conversely, the complex with the potentially hexadentate tpen,
Figure 3. Yields of alkene epoxidations by PAA catalyzed by the
R
bispicen complex 1 as a function of time. The reaction conditions are
identical with those listed for Table 2.
3, is the worst catalyst.
In the previously studied alkene oxidation catalyzed by
[
Ga(phen) Cl ]Cl, the selectivity for the epoxide product is lost
2 2
upon switching to a lower catalyst loading. With a 0.1% catalyst
loading of the phen complex, the most heavily represented
products result from allylic C−H oxidation and bishydrox-
occur on a much shorter time scale. The data show that the
catalytic activity of 1 is maintained past 1 h; the observed
increases in the yields beyond this time point are too large to be
attributed to the uncatalyzed reactions between the olefins and
2
0
ylation. A loss of selectivity for the epoxide is also observed
for the oxidation of cyclohexene catalyzed by 1 but with a
substantially different product distribution. With a 0.1% loading
of 1, 3.0 M cyclohexene is oxidized by 6.0 M PAAR to
predominantly trans-1,2-cyclohexanediol (49%), with cyclo-
hexene oxide as a minor product (6%) and the remainder of the
substrate failing to react (45%). Unexpectedly, 2-cyclohexenol
is not observed. Another potential product of allylic C−H
oxidation, 2-cyclohexenone, is not formed in the olefin
oxidations catalyzed by low concentrations of either 1 or
20
terminal oxidant. The most reactive substrate, cyclooctene, is
fully converted to cyclooctene oxide by 2 h. Beyond 5 h, diol
products consistent with the subsequent opening of the epoxide
are observed. With cyclohexene as the substrate, cyclohexene
oxide is the sole organic product until 5 h, forming in 75% yield
from the alkene. By 6 h, noticeable amounts of trans-1,2-
cyclohexanediol are observed in addition to the epoxide.
MS of the cyclohexene epoxidation at 3 h reveals that the
III
Ga -bispicen adduct is still intact. At both 5 min and 3 h, the
20
[
Ga(phen) Cl ]Cl.
most intense feature has m/z 369.0817, consistent with
2
2
+
Time Scale of Reactivity. The ability of the gallium(III)
[Ga(bispicen-H)(CH CO )] (predicted m/z 369.0842). This
3
2
complexes to catalyze the epoxidation of olefins by PAA was
suggests that the two chloride ions in 1 are displaced by acetate
shortly after the beginning of the reaction. Under ionization
conditions, one of the amine protons is likely lost, explaining
both the reduced mass and the 1+ charge of the observed ion.
The trispicen and tpen complexes 2 and 3 likewise appear to
retain their catalytic activity for longer periods of time than
R
monitored beyond 1 h in order to assess whether a more highly
chelating ligand could extend the lifetime of the reactivity.
Figure 3 shows the yields as a function of time for the
epoxidations of five olefins catalyzed by the bispicen complex 1.
There is no noticeable initiation period, although this may
F
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 4. Yields of alkene epoxidation by PAA catalyzed by 1−6 and [Ga(phen) Cl ]Cl as a function of time. Panel A shows the oxidation of
R
2
2
cyclohexene, whereas B shows the oxidation of 1-octene. The reaction conditions are identical with those listed for Table 2. The data for
[
Ga(phen) Cl ]Cl are from ref 20.
2 2
[
Ga(phen) Cl ]Cl, as assessed by their abilities to catalyze the
complexes with the phen derivatives are likely isolated as
mixtures of cis isomers (Scheme 4).
2
2
oxidation of cyclohexene by PAA . Figure 4A compares the
R
longer term reactivities of compounds 1−6 and [Ga-
Three pyridylamine ligands were investigated: bispicen,
(
phen) Cl ]Cl, using the oxidation of cyclohexene to cyclo-
trispicen, and tpen (Scheme 1). The bispicen molecule is a
2
2
20
35,49−51
hexene oxide as a standard. The plot shows that the activities
of the complexes with the bidentate ligands have decreased
substantially by 3 h. Catalysts 1−3, although slower, appear to
retain more of their activity, with the overall catalytic activity of
well-established tetradentate ligand.
The tpen ligand
typically coordinates metal ions in a hexadentate fashion,
although tetra- and pentadentate modes of coordination have
52−56
been observed.
With respect to previously reported group
1
becoming approximately equivalent to those of 5 and
13 chemistry, there exists a heptacoordinate complex of tpen
III
[
Ga(phen) Cl ]Cl by 3 h. As anticipated, compounds 2 and 3
with In , with the tpen itself providing six of the donor atoms
2
2
57
are much less active than 1 throughout the 3 h. MS analyses of
and 3 at 3 h do not reveal any peaks consistent with the
in the inner sphere. The coordination chemistry of trispicen is
much less established, but a singly methylated derivative, N′-
methyl-N,N,N′-tris(2-pyridylmethyl)-1,2-ethanediamine, is
pentadentate in its structurally characterized complexes with
2
ligand hydrolysis observed in the crystals grown from solutions
of 3. Although 3, for instance, eventually decomposes to 1 and
II
II 37,58
2
(Figure 2), we did not observe m/z features for either of
Mn and Fe .
NMR analysis of 1−3 suggests that all three
III
these compounds in the 3 h data for reactions using 3 as a
catalyst. As with 1, the peaks observed at 3 h are consistent with
acetate adducts.
The plots of the cyclohexene oxide yields versus time for
catalysts 4−6 are more curved than those observed for
compounds 1−3, suggesting that the former set is less durable
under the reaction conditions. The slowing of these three
reactions cannot be attributed to depletion of the substrate,
which only becomes noticeable as the reaction yield exceeds
ligands predominantly coordinate Ga through four N donors,
with one and two pyridine rings failing to ligate the metal in 2
and 3, respectively. This is supported by IR analysis of freshly
precipitated powder samples of these compounds.
Although the crystal structure of 1 (Figure 1) is consistent
with its NMR spectra and elemental analysis, crystals grown
from solutions of 2 are not consistent with the initially obtained
[
(
Ga(trispicen)Cl ]Cl. The crystal structure of the isolated 7
2
Figure 2) demonstrates that the coordination of trispicen to
III
Ga is not strictly limited to the hypodentate mode supported
by the NMR data. In one of the complex ions within the
∼
80% (cyclooctene reaction in Figure 3). Despite this loss of
activity, catalyst 4 remains the best catalyst of the six at the 3 h
time point. Losses of activity for 4 and 5 are also observed over
time when 1-octene is used as the standard (Figure 4B), but
with this particular substrate, the durabilities of 1−3 do not
come close to compensating for their slower activity.
2+
asymmetric unit, [Ga(trispicen)Cl] , the ligand is pentacoor-
−
dinate, whereas in the other complex ion, [GaCl ] , the
4
trispicen ligand is entirely absent from the Ga atom. The
[
Ga(trispicen)Cl]2+ dication represents only the second
III
example of a structurally characterized Ga ion with a N5X
59
coordination sphere (X = F, Cl, Br, I). This coordination
sphere may initially be surprising because the Ga atom does not
obey the octet rule, having five clear Ga−N bonds. The Ga
atom does not strictly obey the octet rule, however, and there
exist an abundance of hexacoordinate gallium(III) species,
including several examples with [GaN6]3+ coordination spheres
DISCUSSION
■
The syntheses of the gallium(III) complexes are generally
straightforward in that no special measures are necessary to
form adducts between the N-donor ligands and the metal ion
and that pure products can be isolated without chromatog-
raphy. The obtained yields range from moderate (51%) to high
60−62
that would seem to deviate even further into hypervalence.
(
85%). One complication with the syntheses is that the
The capability of the trispicen and tpen ligands to more fully
coordinate the Ga ion may explain the decreased epoxidation
activities exhibited by 2 and 3 (Table 2 and Figure 4). Focusing
analysis on the complexes with the three pyridylamine ligands,
III
compounds are hygroscopic, and prolonged exposure to
moisture appears to degrade the pyridylamine ligands, as
evidenced by the crystallized decomposition products of 3. The
G
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
the bispicen complex 1 promotes the most extensive
epoxidation of all substrates except 1-octene, which essentially
does not react with either 1, 2, or 3 present as the catalyst. The
tpen-containing 3 is the worst catalyst of the three compounds,
with 1 h yields that are less than half of those measured for 1.
On the basis of the results, we speculate that the pendant
pyridine rings in 2 and 3 may compete with the terminal
oxidant for the coordination sites vacated by the initially bound
chloride anions. Previous work suggests that the affinities of
alkene substrate. The acetate produced upon O-atom transfer
would also be a better leaving group than the hydroxide or tert-
butyloxide resulting from H O or t-BuOOH.
2
2
1
,10-Phenanthroline is widely viewed as being electron-poor
28−30
relative to amine N-donor ligands
and should render
chelated metal ions more electron-deficient. Complexes 4 and 5
are notable for their ability to catalyze the oxidation of 1-
octene; the other four gallium(III) compounds reported in this
manuscript, conversely, cannot promote this reaction to a
significant degree. The addition of electron-withdrawing nitro
groups onto the phen ligands improves the yields of the
epoxidation reactions; conversely, the installation of electron-
donating amino groups decreases the yields (Table 1).
III
pyridine rings to Ga are comparable to those of carboxylate
3
1
ligands. Reducing the number of pendant pyridine donors
would alleviate the competition between intramolecular and
intermolecular binding and facilitate the coordination of PAA
III
to the Ga center necessary for catalysis. Alternatively, the
Although the complexes with the phen ligands are more
active over shorter durations, the compounds with the
pyridylamine ligands can potentially achieve superior turnover
numbers over longer periods of time because of the greater
stability provided by chelate effects. One noted shortcoming
with the [Ga(phen) Cl ]Cl catalyst is that it decomposed to
steric bulk provided by the additional binding arms in 2 and 3
may hinder coordination of the terminal oxidant and/or limit
the accessibility of the alkene substrate to the reactive portion
of the active oxidant.
The results may suggest that two available coordination sites
are needed to activate PAA most efficiently, which is
reminiscent of the Sharpless-type mechanism proposed for
2
2
unreactive gallium(III) salts and free phen within approximately
3
+
20
the epoxidation of alkenes by mixtures of [Al(H O) ] and
the first hour of the epoxidation reactions. The phen appears
2
6
2
7,63
III
H O .
In the mechanistic scheme proposed by Shul’pin et
al. for this Al catalysis, coordination of both O atoms of H O
to slowly dissociate from the Ga center under acidic
2
2
III
III
2
2
conditions. The stability of the Ga catalyst can be increased
to the metal center precedes a concerted O-atom transfer from
by using a single tetradentate ligand in place of two bidentate
ligands. The ability of bispicen, trispicen, and tpen to sustain
the catalysis over longer periods of time confirms that this is a
27
the bound oxidant to an uncoordinated alkene. If two
coordination sites are likewise needed for the activation of PAA
III
by the Ga centers of the current catalysts, 2 would be
III
viable strategy for Ga -catalyzed olefin epoxidations (Figures 3
anticipated to be an inferior catalyst to 1, in accordance with
our results (Table 2). Scheme 5 illustrates a possible
and 4). Compound 1 is catalytically equivalent to [Ga-
(
phen) Cl ]Cl when the yields of cyclohexene epoxidation
2 2
are measured at 3 h (Figure 4). Unlike the phen compound, the
Ga -bispicen adduct remains intact at 3 h. The observed
Scheme 5
III
decomposition of 2 and 3 (Figures 2 and S1 in the SI) suggests
that the pyridylamine-containing complexes do have finite
catalytic lifetimes. Compounds 1−3 appear to retain their intact
polydentate ligands through the 3 h necessary for the alkene
epoxidation, however, suggesting that the ligand degradation is
slow relative to the alkene epoxidation.
The reactivity with the 1% catalyst loading is selective for the
epoxide product for short durations, with the epoxide
accounting for the entirety of the observed organic product.
The reactions using 1 as the catalyst do start to produce trans-
1,2-cyclohexanediol from cyclohexene between 5 and 6 h after
the start of the reaction. As with the olefin oxidation catalyzed
mechanism for the epoxidation reactions that highlights the
potential benefit of a second coordination site for PAA. The
illustrated mechanism assumes that the N-donor ligands remain
fully bound during the reaction; this assumption has not been
experimentally confirmed. We are currently assessing the
feasibility of various mechanistic pathways through computa-
tional analysis and experiments involving gallium(III) com-
plexes with sterically encumbered bispicen derivatives.
20
by [Ga(phen) Cl ]Cl, the selectivity for the epoxide product
2
2
is lost when the catalyst loading of 1 is reduced to 0.1%. Oddly,
the chemistry associated with 1 differs from that promoted by
the phen compound in that the lower concentration of 1 does
not promote allylic C−H activation. Instead, bishydroxylation is
overwhelmingly favored under such circumstances. When
cyclohexene oxide is used as a substrate with the standard
reaction conditions and a 1% mole loading of gallium (3.0 mM
The higher catalytic activities of the complexes with the phen
III
derivatives suggest that the Ga centers enable catalysis
through their abilities to act as Lewis acids. That more
electron-rich alkenes tend to react more readily would suggest
an electrophilic oxidant, which would be stabilized and
rendered less active by more electron-rich ligands. The
1
, 6.0 M PAA , 0 °C, 7 h), no diol is observed, suggesting that
R
cyclohexene may be converted directly to trans-1,2-cyclo-
hexanediol as opposed to a two-step reaction involving
epoxidation followed by ring opening. We speculate that
III
electrophilicity of the oxidant in the Ga catalysis is also
consistent with the inability of H O and t-BuOOH to serve as
2
2
III
another Ga -based oxidant is responsible and are continuing to
terminal oxidants for the epoxidation reactions. Carbonyl
groups are widely considered to be electron-withdrawing
functional groups. When PAA is used as the terminal oxidant,
its carbonyl group may reduce the electron density on the
transferred O atom sufficiently to allow O-atom transfer to the
investigate this side reactivity through experimental and
computational methods. The mechanism displayed in Scheme
5 certainly does not explain all of the observed oxidative activity
of the gallium(III) complexes.
H
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
(
13) Yamaguchi, M.; Nishimura, Y. Chem. Commun. 2008, 35−48.
14) Wehmschulte, R. J.; Steele, J. M.; Young, J. D.; Khan, M. A. J.
CONCLUSIONS
Six new gallium(III) complexes with N-donor ligands have
been prepared and tested as catalysts for the epoxidation of
■
Am. Chem. Soc. 2003, 125, 1470−1471.
15) Lichtenberg, C.; Spaniol, T. P.; Okuda, J. Inorg. Chem. 2012, 51,
2254−2262.
16) Tang, S.; Monot, J.; El-Hellani, A.; Michelet, B.; Guillot, R.;
(
alkenes by PAA . A comparison of the compounds’ activities
R
III
provides three major insights into Ga -catalyzed alkene
(
epoxidation. First, more electron-deficient ligands are found
to support more rapid catalytic turnover, with phen derivatives
leading to superior initial activity over ligands with amine and
pyridine chelating groups. The gallium(III) complexes with
phen derivatives are markedly better at activating the electron-
deficient substrate 1-octene, and [Ga(NO -phen) Cl ]Cl is the
best catalyst of the six under all characterized reaction
conditions. Second, more highly coordinating ligands, such as
the tetradentate bispicen, can prolong the catalytic activity by
stabilizing the ligand−metal adduct. Third, it is found that the
more highly chelating N-donor ligands, trispicen and tpen,
decrease the catalytic activity. The additional binding arms may
compete with terminal oxidant for the coordination sites, or
they may impede the reactivity through steric effects.
Bour, C.; Gandon, V. Chem.Eur. J. 2012, 18, 10239−10243.
(17) Carty, A. J.; Dymock, K. R.; Boorman, P. M. Can. J. Chem. 1970,
48, 3524−3529.
(18) Ivanov-Emin, B. N.; Nisel′son, L. A.; Rabovik, Y. I.; Larionova,
L. E. Russ. J. Inorg. Chem. 1961, 6, 1142−1146.
(19) McPhail, A. T.; Miller, R. W.; Pitt, C. G.; Gupta, G.; Srivastava,
2
2
2
S. C. J. Chem. Soc., Dalton Trans. 1976, 1657−1661.
20) Jiang, W.; Gorden, J. D.; Goldsmith, C. R. Inorg. Chem. 2012,
1, 2725−2727.
21) Fricke, R.; Kosslick, H.; Lischke, G.; Richter, M. Chem. Rev.
000, 100, 2303−2405.
22) Pescarmona, P. P.; Janssen, K. P. F.; Jacobs, P. A. Chem.Eur. J.
007, 13, 6562−6572.
23) Stoica, G.; Santiago, M.; Jacobs, P. A.; Per
Pescarmona, P. P. Appl. Catal., A 2009, 371, 43−53.
24) Cates, C. D.; Myers, T. W.; Berben, L. A. Inorg. Chem. 2012, 51,
1891−11897.
25) Nam, W.; Valentine, J. S. J. Am. Chem. Soc. 1990, 112, 4977−
979.
(26) Rinaldi, R.; de Oliveira, H. F. N.; Schumann, H.; Schuchardt, U.
J. Mol. Catal. A 2009, 307, 1−8.
27) Kuznetsov, M. L.; Kozlov, Y. N.; Mandelli, D.; Pombeiro, A. J.
L.; Shul’pin, G. B. Inorg. Chem. 2011, 50, 3996−4005.
28) Berryman, O. B.; Bryantsev, V. S.; Stay, D. P.; Johnson, D. W.;
Hay, B. P. J. Am. Chem. Soc. 2007, 129, 48−58.
29) Ishi-i, T.; Hirashima, R.; Tsutsumi, N.; Amemori, S.; Matsuki, S.;
Teshima, Y.; Kuwahara, R.; Mataka, S. J. Org. Chem. 2010, 75, 6858−
868.
30) Oguadinma, P. O.; Rodrigue-Witchel, A.; Reber, C.; Schaper, F.
Dalton Trans. 2010, 39, 8759−8768.
31) Caravan, P.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1997, 36, 1306−
315.
32) Murphy, A.; Pace, A.; Stack, T. D. P. Org. Lett. 2004, 6, 3119−
122.
33) Sheldrick, G. M. SHELXTL, 6.12 ed.; Siemens Analytical X-ray
(
5
(
2
(
2
(
́
ez-Ramírez, J.;
(
1
(
4
ASSOCIATED CONTENT
Supporting Information
Crystallographic data for [Ga(trispicen)Cl]Cl , which forms
■
*
S
2
1
13
upon decomposition of 3, mass spectrometry of 3, H and C
NMR spectra for 1−6; IR spectra for 1−7, figure showing 1,2-
cyclohexanediol production over 7 h, mass spectrometry of the
gallium adducts in 1−3-catalyzed epoxidations of cyclohexene
at 3 h, and CIF file of CCDC reference numbers 913806−
(
(
9
13808. This material is available free of charge via the Internet
(
at http://pubs.acs.org.
6
(
AUTHOR INFORMATION
Corresponding Author
E-mail: crgoldsmith@auburn.edu.
■
(
1
(
3
*
Notes
The authors declare no competing financial interest.
(
ACKNOWLEDGMENTS
■
Instruments, Inc.: Madison, WI, 2001.
(34) Sheldrick, G. M. SADABS; Bruker Analytical X-ray Systems:
Madison, WI, 1996.
The authors are grateful for funding provided by Auburn
University and a Doctoral New Investigator grant from the
Petroleum Research Fund, sponsored by American Chemical
Society (Grant #49532-DNI3).
(35) Toftlund, H.; Pedersen, E.; Yde-Andersen, S. Acta Chem. Scand.
A 1984, 38, 693−697.
(
(
36) Anderegg, G.; Wenk, F. Helv. Chim. Acta 1967, 50, 2330−2332.
́
37) Mialane, P.; Nivorojkine, A.; Pratviel, G.; Azema, L.; Slany, M.;
REFERENCES
■
Godde, F.; Simaan, A.; Banse, F.; Kargar-Grisel, T.; Bouchoux, G.;
Sainton, J.; Horner, O.; Guilhem, J.; Tchertanova, L.; Meunier, B.;
Girerd, J.-J. Inorg. Chem. 1999, 38, 1085−1092.
(
1) Chi, Y.; Chou, T.-Y.; Wang, Y.-J.; Huang, S.-F.; Carty, A. J.;
Scoles, L.; Udachin, K. A.; Peng, S.-M.; Lee, G.-H. Organometallics
2
(
004, 23, 95−103.
(
38) Goldsmith, C. R.; Jonas, R. T.; Cole, A. P.; Stack, T. D. P. Inorg.
Chem. 2002, 41, 4642−4652.
39) Junk, P. C.; Skelton, B. W.; White, A. H. Aust. J. Chem. 2006, 59,
47−154.
40) Gill, N. S.; Nuttall, R. H.; Scaife, D. E.; Sharp, D. W. A. J. Inorg.
Nucl. Chem. 1961, 18, 79−87.
41) Willett, R. D.; Gomez-García, C. J.; Twamley, B.; Gom
S.; Ruiz, E. Inorg. Chem. 2012, 51, 5487−5493.
42) Goodson, P. A.; Glerup, J.; Hodgson, D. J.; Michelsen, K.;
Pedersen, E. Inorg. Chem. 1990, 29, 503−508.
43) Coates, C. M.; Fiedler, S. R.; McCullough, T. L.; Albrecht-
2) Park, J.-H.; Afzaal, M.; Helliwell, M.; Malik, M. A.; O’Brien, P.;
Raftery, J. Chem. Mater. 2003, 15, 4205−4210.
(
1
(
(3) Valet, M.; Hoffman, D. M. Chem. Mater. 2001, 13, 2135−2143.
(4) Suh, S.; Hoffman, D. M. Chem. Mater. 2000, 12, 2794−2797.
(5) MacInnes, A. N.; Power, M. B.; Barron, A. R. Chem. Mater. 1993,
5
(
(
(
(
, 1344−1351.
(
́
́
ez-Coca,
6) Kouvetakis, J.; Beach, D. B. Chem. Mater. 1989, 1, 476−478.
7) Cowley, A. H. J. Organomet. Chem. 2000, 600, 168−173.
(
̈
8) Bartholoma, M. D. Inorg. Chim. Acta 2012, 389, 36−51.
9) Wadas, T. J.; Wong, E. H.; Weisman, G. R.; Anderson, C. J.
(
Chem. Rev. 2010, 110, 2858−2902.
10) Bartholoma, M. D.; Louie, A. S.; Valliant, J. F.; Zubieta, J. Chem.
Rev. 2010, 110, 2903−2920.
11) Boros, E.; Ferreira, C. L.; Cawthray, J. F.; Price, E. W.; Patrick,
B. O.; Wester, D. W.; Adam, M. J.; Orvig, C. J. Am. Chem. Soc. 2010,
32, 15726−15733.
12) Dagorne, S.; Atwood, D. A. Chem. Rev. 2008, 108, 4037−4071.
(
̈
Schmitt, T. E.; Shores, M. P.; Goldsmith, C. R. Inorg. Chem. 2010, 49,
1481−1486.
(44) Collins, M. A.; Hodgson, D. J.; Michelsen, K.; Towle, D. K. J.
Chem. Soc., Chem. Commun. 1987, 1659−1660.
(45) Arulsamy, N.; Hodgson, D. J.; Glerup, J. Inorg. Chim. Acta 1993,
209, 61−69.
(
1
(
I
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
46) Ardon, M.; Bino, A.; Michelsen, K.; Pedersen, E.; Thompson, R.
C. Inorg. Chem. 1997, 36, 4147−4150.
47) Restivo, R.; Palenik, G. J. J. Chem. Soc., Dalton Trans. 1972,
41−344.
48) Sofetis, A.; Raptopoulou, C. P.; Terzis, A.; Zafiropoulos, T. F.
Inorg. Chim. Acta 2006, 359, 3389−3395.
49) Glerup, J.; Goodson, P. A.; Hazell, A.; Hazell, R.; Hodgson, D.
(
3
(
(
J.; McKenzie, C. J.; Michelsen, K.; Rychlewska, U.; Toftlund, H. Inorg.
Chem. 1994, 33, 4105−4111.
(
50) Arulsamy, N.; Goodson, P. A.; Hodgson, D. J.; Glerup, J.;
Michelsen, K. Inorg. Chim. Acta 1994, 216, 21−29.
51) Heinrichs, M. A.; Hodgson, D. J.; Michelsen, K.; Pedersen, E.
Inorg. Chem. 1984, 23, 3174−3180.
52) Hureau, C.; Blanchard, S.; Nierlich, M.; Blain, G.; Rivier
Girerd, J.-J.; Anxolabehere-Mallart, E.; Blondin, G. Inorg. Chem. 2004,
3, 4415−4426.
53) Musso, S.; Anderegg, G.; Ruegger, H.; Schlap
Gramlich, V. Inorg. Chem. 1995, 34, 3329−3338.
54) Chang, H.-R.; McCusker, J. K.; Toftlund, H.; Wilson, S. R.;
Trautwein, A. X.; Winkler, H.; Hendrickson, D. N. J. Am. Chem. Soc.
990, 112, 6814−6827.
55) Yoon, D. C.; Lee, U.; Oh, C. E. Bull. Korean Chem. Soc. 2004,
5, 796−800.
56) Dueland, L.; Hazell, R.; McKenzie, C. J.; Nielsen, L. P.;
Toftlund, H. J. Chem. Soc., Dalton Trans. 2001, 152−156.
57) Harrington, J. M.; Oscarson, K. A.; Jones, S. B.; Reibenspies, J.
H.; Bartolotti, L. J.; Hancock, R. D. Z. Naturforsch. 2007, 62b, 386−
96.
58) Hureau, C.; Groni, S.; Guillot, R.; Blondin, G.; Duboc, C.;
Anxolabehere-Mallart, E. Inorg. Chem. 2008, 47, 9238−9247.
59) Radish, K. M.; Cornillon, J.-L.; Korp, J. D.; Guilard, R. J.
Heterocycl. Chem. 1989, 26, 1101−1104.
60) Zhang, Q.; Malliakas, C. D.; Kanatzidis, M. G. Inorg. Chem.
009, 48, 10910−10912.
61) Hilfiker, K. A.; Brechbiel, M. W.; Rogers, R. D.; Planalp, R. P.
Inorg. Chem. 1997, 36, 4600−4603.
62) Baker, R. J.; Jones, C.; Kloth, M.; Mills, D. P. New J. Chem.
004, 28, 207−213.
63) Sharpless, K. B.; Townsend, J. M.; Williams, D. R. J. Am. Chem.
Soc. 1972, 94, 295−296.
(
(
̀
e, E.;
́
̀
4
(
̈
fer, C. W.;
(
1
(
2
(
(
3
(
́
̀
(
(
2
(
(
2
(
J
dx.doi.org/10.1021/ic400570h | Inorg. Chem. XXXX, XXX, XXX−XXX