Alkane oxidation reactivity of homogeneous and heterogeneous metal complex catalysts with mesoporous silica-immobilized (2-pyridylmethyl)amine type ligands

Article abstract of DOI:10.1016/j.mcat.2017.09.027

Attachment of N,N-bis(2-pyridylmethyl)amine (L1) and N-(2-pyridylmethyl)glycine (L2) ligands on the azide-functionalized SBA-15 type mesoporous silica support and subsequent insertion of metal salts into the ligand-immobilized supports yielded heterogeneous metal complex catalysts, M(A)/SBA*-Ln-x, where M, A, *, n, and x indicate the metal ion (Mn^II, Fe^II, Co^II, Ni^II, Cu^II), counter anion (Cl, OAc, OTf), trimethylsilyl end-capped silica, the numbering of the ligands, and the initial content of the azide tether group (tether/Si mol%) respectively. The cyclohexane oxidation activity of these immobilized catalysts and corresponding homogeneous complexes have been evaluated with the use of m-chloroperbenzoic acid as an oxidant. The activity of the immobilized catalysts is influenced by the ligand density on the support. Both the site-isolated immobilized catalysts and the related homogeneous catalysts of ligand L1 show similar trends of their final TONs on the type of metal ions (Co ≈ Ni > Fe > Mn > Cu). These results suggest that all of the metal complex sites are successfully immobilized without structural changes under the site-isolated condition. The reactivity trends of the complexes with the L2 ligand and their dependence on the ligand density were complicated. The observed results may be explained by the formation of cluster complexes via bridging carboxylate moieties of the ligand.

Full text of DOI:10.1016/j.mcat.2017.09.027

Molecular Catalysis 443 (2017) 14–24
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
Editor’s choice paper
Alkane oxidation reactivity of homogeneous and heterogeneous metal
complex catalysts with mesoporous silica-immobilized
(2-pyridylmethyl)amine type ligands
Jun Nakazawa∗, Yuma Doi, Shiro Hikichi∗
Department of Material and Life Chemistry, Faculty of Engineering, Kanagawa University, 3-27-1 Rokkakubashi, Kanagawa-ku, Yokohama 221-8686, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Attachment of N,N-bis(2-pyridylmethyl)amine (L1) and N-(2-pyridylmethyl)glycine (L2) ligands on the
azide-functionalized SBA-15 type mesoporous silica support and subsequent insertion of metal salts into
the ligand-immobilized supports yielded heterogeneous metal complex catalysts, M(A)/SBA*-Ln-x, where
M, A, *, n, and x indicate the metal ion (MnII, FeII, CoII, NiII, CuII), counter anion (Cl, OAc, OTf), trimethylsilyl
end-capped silica, the numbering of the ligands, and the initial content of the azide tether group (tether/Si
mol%) respectively. The cyclohexane oxidation activity of these immobilized catalysts and corresponding
homogeneous complexes have been evaluated with the use of m-chloroperbenzoic acid as an oxidant.
The activity of the immobilized catalysts is influenced by the ligand density on the support. Both the site-
isolated immobilized catalysts and the related homogeneous catalysts of ligand L1 show similar trends
of their final TONs on the type of metal ions (Co ≈ Ni > Fe > Mn > Cu). These results suggest that all of
the metal complex sites are successfully immobilized without structural changes under the site-isolated
condition. The reactivity trends of the complexes with the L2 ligand and their dependence on the ligand
density were complicated. The observed results may be explained by the formation of cluster complexes
via bridging carboxylate moieties of the ligand.
Received 2 August 2017
Received in revised form
21 September 2017
Accepted 23 September 2017
Keywords:
Immobilized catalyst
Metal complex
Alkane oxidation catalysis
Peracid
Site density
© 2017 Elsevier B.V. All rights reserved.
1. Introduction
on the metal sites and subsequent C H bond scissions have con-
tributed toward our understanding of metalloenzymes in the field
Owing to its high reactivity toward organic molecules, meta-
chloroperbenzoic acid (mCPBA) has been widely applied as an
oxidant for laboratory level syntheses [1]. This oxidant can react
with inactive saturated hydrocarbon molecules such as cyclohex-
ane by gentle heating over 60 ◦C. For this reaction, metal salts or
metal complexes act as catalysts to accelerate the reaction rate
and to change product selectivity via the coordination and acti-
vation of mCPBA on the metal center [2]. Significant research effort
has been invested in the development of metal complex catalysts
for selective oxidation reactions with the control of site selec-
of bioinorganic chemistry [6,7].
In 2006, Itoh and co-workers found that a nickel(II) tpa complex
(tpa = tris(2-pyridylmethyl)amine ligand) showed high catalytic
activity for the oxidation of cyclohexane at ambient temperature
[8]. In their report, these authors described a series of metal com-
plexes of divalent metal ions (M = Mn, Fe, Co, Ni) and tpa ligand:
[(tpa)Mn-(␮2-1,3-OAc)2-Mn(tpa)](BPh4)2, [(tpa)Fe-(␮-1,3-OAc)2-
Fe(tpa)](BPh4)2, [Co(␬2-OAc)(tpa)](BPh4)2, and [Ni(␬1-OAc)(H2O)
(tpa)](BPh4). The metal complexes described in this article
are denoted according to the scheme ‘metal and (anions) in
source/ligand’. Therefore, the above mentioned complexes can
be written in the abbreviated form as MnII(OAc,BPh4)/tpa,
FeII(OAc,BPh4)/tpa, CoII(OAc,BPh4)/tpa, and NiII(OAc,BPh4)/tpa.
Among these catalysts, the difference in the metal ion caused a
change in the reactivity, which is evaluated by the catalyst turnover
number (TON, NiII > FeII > CoII > MnII) and the product selectiv-
ity for alcohol over ketone (A/K selectivity, CoII > NiII > FeII » MnII).
The catalyst NiII(OAc,BPh4)/tpa has inspired further investigations
by other research groups into nickel complexes for reactiv-
tivity (e.g. 1◦–3◦
C H, steric, and electronic) and over oxidation
(e.g. alcohols to ketones/aldehydes to carboxylates/lactones) [3–5].
Furthermore, mechanistic studies of the activation of the oxidant
∗
Corresponding authors at: Department of Material and Life Chemistry, Faculty of
Engineering, Kanagawa University, 3-27-1 Rokkakubashi, Kanagawa-ku, Yokohama
221-8686, Japan.
E-mail addresses: jnaka@kanagawa-u.ac.jp (J. Nakazawa),
hikichi@kanagawa-u.ac.jp (S. Hikichi).
http://dx.doi.org/10.1016/j.mcat.2017.09.027
2468-8231/© 2017 Elsevier B.V. All rights reserved.

J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
15
support. Although the “ligand isolated” immobilized nickel com-
plex catalyst (NiII/SBA*-L1-0.5) was highly reactive, this catalyst
showed metal leaching into the reaction solution during catalysis.
In the case of the cobalt(II) complex with the L1 ligand, the “ligand
dense” catalyst (CoII/SBA*-L1-4) showed improved product selec-
tivity for alcohol over ketone (A/K) ratio than that for free cobalt
ions. Due to the free cobalt ions, faster catalysis was observed with
low A/K values and the suppression of metal leaching at the “ligand
dense” sites increased the reactivity of the CoII/L1 site.
In this article, we have explored three factors, namely, (1) type of
metal source, (2) type of ligand, and (3) immobilization of the metal
complexes with regulated site density, in order to understand the
general features of homogeneous and immobilized complexes as
well as to develop efficient catalysts (Scheme 1). We first synthe-
sized complexes of metal ions such as MnII, FeII, FeIII, and CuII ions
in addition to previously reported NiII and CoII ions for the het-
erogeneous SBA*-L1-x support. Second, we synthesized complexes
with the new ligand, N-(2-pyridylmethyl)glycine (L2), which has a
carboxylate side arm. The anionic O-donor in the side arm of this
ligand shows enhanced affinity especially toward high valent Mn
and Fe ions compared to L1 and helps prevent metal leaching during
catalysis. Third, we investigated the influence of the surface ligand
density of the immobilized complex on the catalytic activity. In
addition, we were also interested in the development of immobi-
lized catalysts of the L2 ligand from the viewpoint of the biomimetic
chemistry of nonheme iron enzymes in which the iron centers are
supported by imidazole and carboxylate donors from residues of
their protein backbone [32–35].
Scheme 1. Preparation of the immobilized metal complex catalyst M(A)/SBA*-Ln-x
on a mesoporous silica support, where x denotes the azide content of the silica, Ln
indicates the numbering of the ligands, * denotes the TMS end-capped support, and
M(A) is the general description of metals and anions of the applied metal source.
2. Experimental
2.1. Materials and physico-chemical characterization techniques
ity enhancement, characterization of reaction intermediates, and
understanding the reaction mechanism [9–16].
Atomic absorption analysis was performed on a Shimadzu
AA-6200 instrument. Elemental analysis was performed on a
Perkin-Elmer CHNS/O analyzer 2400II. ESI- and CSI-MS spec-
tra were measured on a JEOL JMS-T100LC mass spectrometer.
GC analysis was performed on a Shimadzu GC-2010 gas chro-
matograph with a flame ionization detector (FID) and an Rtx-5
column (Restek, 30 m length, 0.25 mm i.d., and 0.25 ␮m thickness).
IR spectra were recorded on a JASCO FT/IR 4200 spectrometer.
NMR spectra were recorded on a JEOL ECA-500 spectrometer.
UV–vis spectra were measured on a JASCO V650 spectrom-
eter with a PIN-757 integrating sphere attachment for solid
reflectance. Nitrogen sorption studies were performed at the liq-
uid nitrogen temperature of 77 K using Micromeritics TriStar
3000. Before the adsorption experiments, the samples were out-
gassed under reduced pressure for 3 h at 333 K. Single crystal
X-ray diffraction data was collected using a Rigaku Saturn 70
CDD area detector system with graphite-monochromated Mo-
K␣ radiation. Crystal information files (CIF) of the complexes
reported in this paper have been deposited with the Cam-
bridge Crystallographic Data Centre as supplementary publications
CCDC1536605 [FeIIICl3(L1)], CCDC1536612 [FeIIICl2(L1)](BPh4),
and CCDC1536613 [CuII(OAc)(L1)](BF4)·(Et2O)0.5. This data can be
obtained online free of charge from the Cambridge Crystallographic
Data Centre at www.ccdc.cam.ac.uk/data request/cif.
Consequently, several groups, including ours, have investi-
gated the immobilization of these complexes on silica supports to
develop further catalytic applications and to obtain any information
on the reaction mechanism [17–21]. To date, several prepara-
tions, characterization, and reactivity studies of mesoporous silica
supported metal complex catalysts with tpa and the N,N-bis(2-
pyridylmethyl)-N-{(1H-1,2,3-triazol-4-yl)methyl}amine (L1) lig-
and (Scheme 1) have been reported [22–25]. Mesoporous materials
have attracted much attention due to their potential for various
applications including catalyst support [26–28]. The mesoporous
silica support [29], SBA-15, allows an easy access of the reagents to
the reaction sites in its mesoscale channels and minimizes the inter-
site interactions because of its large surface area (∼700 m2 g−1
)
and pore diameter (∼5 nm). Usage of L1 ligand simplifies the cat-
alyst preparation with a tpa-type ligand [30,31]. We investigated
the reactivities of homogeneous as well as immobilized nickel(II)
and cobalt(II) complex catalysts with L1 ligand (metal/SBA*-L1-
x, where * indicated the trimethylsilyl treatment of the support
surface and x = tether/Si mol%) for the catalytic oxidation of cyclo-
hexane by mCPBA [24,25].
In our previous research on immobilized nickel and cobalt cat-
alysts with SBA*-L1-x supports, we found that the “surface ligand
density” on the supports effected the catalytic activity even though
the trend of reactivity was opposite for the two metals. In the
case of the Ni catalyst, the reactivity decreased with increasing
surface ligand density (x = 0.5 to 4 mol%). Based on the spectro-
scopic comparison with the homogeneous model complexes, we
concluded that the loss in activity of the immobilized catalyst orig-
inated from the [Ni(L1)(X)n] site (X: labile coordinating solvent or
ligand) with increase in the coordinatively saturated and catalyt-
ically inert [Ni(L1)2] site under “ligand dense” conditions on the
All commercial reagents and solvents were used as
received without further purification unless otherwise noted.
meta-Chloroperbenzoic acid (mCPBA) was washed with KH2PO4-
Na2HPO4 buffer solution (pH 7.4) and water in order to remove
meta-chlorobenzoic acid (mCBA). Azidopropyl-functionalized
mesoporous silica SBA-N3-x (x = N3/Si mol%) was prepared accord-
ing to our previously reported method [23–25]. The ligand
immobilized and trimethylsilyl end-capped mesoporous silicas

16
J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
Scheme 4. Preparation of homogeneous metal complexes with L2.
capped support. The corresponding homogeneous catalysts have
been described as M(A)/Ln in a similar manner.
Scheme 2. Preparation of L2.
2.2. Preparation of immobilized complex catalysts
2.2.1. Immobilization of L2 on SBA-N3-0.5 (General procedure of
ligand attachment)
A mixture of SBA-N3-0.5 (1.5 g), ligand precursor 3 (333 ␮mol;
2 equiv. for N3 on the silica support), and CuIISO4·5H2O (417.5 mg,
1.67 mmol; 10 equiv. for N3) was suspended in a solvent mix-
ture of DMF and H2O (40 mL each) under Ar atmosphere and
stirred for 30 min. To this suspension, sodium ascorbate solution
(136 mg, 0.687 mmol in 5 mL of H2O; 2 equiv. for Cu) was added
and the mixture was stirred at 80 ◦C for 24 h. After the reaction,
the silica support was recovered by filtration and washed with
DMF and water. In order to remove Cu ions, silica was washed
with a methanol solution of Et2NCS2Na (0.1 M), DMF, methanol,
and acetone. This Cu removal process was repeated until the fil-
trate become colorless. Finally, rinsing with a methanol solution
of sodium acetate (0.01 M) and subsequent vacuum drying yielded
a pale yellow powder (1.4 g). Ligand immobilization on the other
SBA-N3-x silicas (x = 1, 2, 4%) were also performed under similar
conditions.
2.2.2. TMS end-capping of SBA-L2-0.5 silica (General procedure
of surface passivation)
SBA-L2-0.5 (1.4 g) was suspended in dry toluene (50 mL) and
1,1,1,3,3,3-hexamethyldisilazane (0.8 mL, 3.8 mmol) was added
under Ar atmosphere. The mixture was stirred for 1 h at 40 ◦C and
subsequently, the silica support was recovered by filtration and
washed with toluene. Finally, vacuum drying yielded the TMS end-
capped silica SBA*-L2-0.5 silica as a pale yellow powder (1.4 g). The
end-capping on the other SBA-L2-x silicas (x = 1, 2, 4%) was also
performed under the same conditions.
Scheme 3. Preparation of homogeneous metal complexes with L1.
SBA*-L1-x were prepared following our previously reported
procedure [24,25]. For the homogeneous analogue, N,N-bis(2-
pyridylmethyl)-N-{(1-tert-butyl-1H-1,2,3-triazol-4-yl)methyl}
amine (L1) was also prepared according to our previous report
[24,25]. The synthesis of L2 (Scheme 2) and the preparation
of homogeneous complex catalysts with L1 and L2 ligands
(Schemes 3 and 4) are described in the supporting information.
The immobilized metal complex catalysts described in this work
have been denoted as M(A)/SBA*-Ln-x, where M(A) refers to the
metals and anions of the applied metal source, Ln is the num-
bering of the ligands, x denotes the content of azide functionality
mixed during the silica synthesis, and * indicates the TMS end-
2.2.3. Synthesis of CuII(OAc)/SBA*-L1-4 (General procedure of
metal incorporation)
After 0.205 g of SBA*-L1-4 was suspended into a methanol
solution of Cu(OAc)2 (∼1.2 eq. for the ligand, 50 mL) under Ar atmo-
sphere, the mixture was stirred for 1 h. The silica powder was
recovered by filtration and then washed with methanol. Subse-

J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
17
Table 1
Ligand and metal loadings on M(A)/SBA*-L1-x catalysts at before and after usage for the catalysis.
Metal sourceand solvent
xa
L1 (mmol/g)
Metal (mmol/g)before/after usage
L1/M ratio
Leached metal (%)b
MnCl2
in MeOH
0.5
4
0.13
0.27
0.11/0.04
0.18/0.02
1.2
1.5
64
89
Fe(OTf)2
in MeCN
0.5
1
2
0.08
0.10
0.18
0.37
0.07/0.04
0.10/— d
0.15/— d
0.25/0.09
1.1
1.0
1.2
1.5
43
d
—
—
d
4
64
c
Co(OAc)2
in MeOH
in MeOH
0.5
4
0.08
0.32
0.08/0.03
0.23/0.16
1.0
1.4
67
30
c
Ni(OAc)2
0.5
1
2
0.07
0.12
0.25
0.33
0.08/0.03
0.14/0.05
0.22/0.14
0.19/0.17
0.9
0.9
1.1
1.7
63
64
36
11
4
Cu(OAc)2
in MeOH
in MeOH
0.5
4
0.089
0.29
0.09/0.08
0.29/0.31
1.0
1.0
13
0
d
Zn(OAc)2
0.5
4
0.089
0.28
0.07/—d
0.15/—d
1.2
1.9
—
—
d
a
(N3 ∼ Si)/Si mol% in the silica preparation.
After usage for the catalysis.
Obtained from reference [24,25].
No data.
b
c
d
Table 2
Ligand and metal loadings on M(A)/SBA*-L2-x catalysts before and after the catalysis.
Metal sourceand solvent
xa
L2 (mmol/g)
Metal (mmol/g)before/after usage
L2/M ratio
Leached metal (%)b
MnCl2
in MeCN
0.5
4
0.04
0.32
0.04/0.01
0.28/0.05
1.0
1.1
83
82
Fe(OTf)2
in MeCN
0.5
1
2
0.05
0.10
0.18
0.32
0.05/0.04
0.11/— c
0.16/— c
0.25/0.17
1.0
0.9
1.1
1.3
20
c
—
—
c
4
32
Co(OAc)2
In MeCN
In MeCN
0.5
4
0.04
0.34
0.05/0.03
0.32/0.15
1.3
1.1
40
30
Ni(OAc)2
0.5
4
0.04
0.34
0.04/0.03
0.27/0.21
1.0
1.3
25
22
a
(N3 ∼ Si)/Si mol% in the silica preparation.
After usage for the catalysis.
No data.
b
c
quently, it was dried under reduced pressure at room temperature
for 24 h to give CuII(OAc)/SBA*-L1-4 (0.210 g). The different metal
sources and solvents applied for the incorporation of the metal
into the ligand immobilized silicas SBA*-Ln-x are summarized in
Tables 1 and 2.
General procedure for determining the metal loadings of
M/SBA*-Ln-x (n = 1 or 2): M/SBA*-Ln-x (5 mg) was dissolved into
1 mL of aq. 10% potassium hydroxide by heating for 5 min. The
solution was then acidified with 2 mL of conc. aq. nitric acid and
diluted with H2O to a total volume of 10 mL. The solution was
passed through a syringe filter prior to its analysis using an atomic
absorption spectrometer.
2.3. Determination of ligand and metal contents on silica
The loading of the ligand L1 in SBA*-L1-x was quantified by
UV–vis absorption spectroscopy. SBA*-L1-x (5 mg) and acidic NH4F
(18 mg) were dissolved in 5 mL of MeOH with gentle heating (50 ◦C)
for several hours. The solution was diluted with methanol and conc.
aq. HCl (ca. 1 mL) to a volume of 100 mL and the UV–vis spectrum
of the solution was measured in a 1 cm cell. The ligand loading was
calculated by using the absorption coefficient of the homogeneous
analogue L1 (ꢀ = 263 nm, ε = 11700 mol−1 dm3 cm−1 in aq. HCl) as a
reference.
2.4. Catalytic oxidation of cyclohexane
The catalytic oxidation of cyclohexane was carried out at
room temperature under Ar atmosphere, except in the case of
cobalt complexes (35 ◦C). In a typical reaction, a solution (or sus-
pension) of the catalyst (2.0 ␮mol of the metal ion dissolved
in 1.0 mL of acetonitrile and dichloromethane each) was added
to a mixture of cyclohexane (15 mmol, 1.6 mL), mCPBA oxidant
(345 mg, 2.0 mmol), and nitrobenzene (10 ␮L, 100 ␮mol, used as
GC standard) in dichloromethane (2.0 mL). In order to monitor the
formation of the products, 0.2 mL of the reaction mixture was col-
lected after a certain period and quenched with a dichloromethane
solution (0.5 mL) of triphenylphosphine (10 mg), after which the
solution was subjected to GC analysis. The averages of three inde-
pendent reaction measurements have been reported. No change in
The loading of ligand L2 was quantified by 1H NMR. SBA*-L2-
x (10.0 mg) and a pellet of potassium hydroxide were dissolved in
1 mL of D2O with heating for several minutes. Subsequently, a solu-
tion of p-nitrophenol (3.6 ␮mol in 0.10 mL of D2O) was added as an
internal standard. The loading amount of the ligand was calculated
by the integration ratio of the pyridyl protons of the ligand to the
aromatic protons of p-nitrophenol (Fig. S6).

18
J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
for L1, with the exception of CuII(OAc)/SBA*-L1-4. Partial forma-
tion of [ML2] species may be involved in the case of MnII(Cl)-,
FeII(OTf)/SBA*-L1-4, as previously reported for the NiII/SBA*-L1-x
system [24]. The structural changes to the [Ni(L1)2] species occur
on the surface of the NiII/SBA*-L1-x system, as indicated by the
change in diffuse reflectance UV–vis spectra. The conversion of
[ZnII(L1Ar)]2+ and [ZnII(L1Ar)2]2+ (Ar indicates replacement to an
aromatic group instead of the tBu group on L1) [36] in homoge-
neous systems has also been reported. According to the molecular
structure of the related homogeneous CuII(OAc,BF4)/L1 complex
mentioned in Section 3.2.1, the low L1/Cu ratio (1.0) in CuII/SBA*-
L1-4 may be explained by the following two effects; (1) the square
planer geometry of the CuII center reduces the bulkiness of the com-
plex, and/or (2) the inter molecular coordination of the triazole side
arm forms a tightly packed chain structure.
reactivity was observed when the reaction was shaken instead of
mechanical stirring with a Teflon-coated magnetic bar.
3. Results and discussion
3.1. Preparation and characterization of immobilized catalysts
In this study, we extended the variety of metal ions from pre-
viously reported NiII and CoII ions to MnII, FeII, FeIII, and CuII ions
to be used for the synthesis of heterogeneous SBA*-L1-x systems.
Furthermore, a new ligand, L2, with an anionic carboxylate side
arm was utilized. This ligand exhibits an enhanced affinity for man-
ganese and iron ions compared to L1 and also helps in preventing
metal leaching during catalysis.
The ligand immobilized silicas, SBA*-Ln-x, were prepared
according to the previously reported procedure (shown in
Schemes 1 and 2) [24,25]. The preparation and immobilization pro-
cedure of the L2 system are shown in Scheme 2. Initially, SBA-N3-x
was prepared from Si(OEt)4 and 3-azidepropyltriethoxysilane in
the presence of P123 template micelle under acidic condition. In this
process, the content of the azide tether in the silica precursors was
regulated to x mol% based on the total Si content. The ligands L1 and
L2 were immobilized onto the solid support by the “click reaction”
of the azide tether and alkyne moiety of the ligands with CuI cata-
lyst to give SBA-Ln-x. The decrease of the azide peak at 2110 cm−1
in the IR spectrum of SBA*-L2-x confirmed the progress of the
attachment of L2. Subsequently, SBA*-Ln-x were obtained after
TMS end-capping treatment with 1,1,1,3,3,3-hexamethyldisilazane
to protect the surface silanol site, which was potential donor for
metal ions. The quantification of the ligand L2 on the SBA*-L2-x sup-
port was performed by 1H NMR with p-nitrophenol (as the internal
standard) after decomposition of the support by NaOH/D2O solu-
tion. The ligand loading amounts of L1 and L2 in the SBA*-Ln-x
silicas are given in Tables 1 and 2. In both ligands, the ligand load-
ings were clearly regulated by the initial content of the azide tether
on the support.
The metal ions were incorporated on the ligand-immobilized
supports. The SBA*-L1-x and SBA*-L2-x supports were treated with
methanol or acetonitrile solutions of various metal salts to obtain
M(A)/SBA*-Ln-x supports, where M(A) denotes the applied metal
ions and counter anions. The treatments with iron(II) acetate and
manganese(II) acetate showed low reproducibility of the loading
amount of these metal ions. Therefore, we applied iron(II) tri-
flate and manganese(II) chloride as the alternative metal sources,
respectively. For SBA*-L1-x support, we also applied ZnII acetate
to investigate the metal coordination property of the L1 ligand
sites on the silica support. Atomic absorption spectroscopy after
silica digestion confirmed the loading amounts of the metal ions
and these were compared with each other and with the pre-
viously reported NiII(OAc)/SBA*-L1-x (x = 0.5, 1, 2, 4 mol%) and
CoII(OAc)/SBA*-L1-x (x = 0.5 and 4 mol%) catalysts (Tables 1 and 2)
[24,25].
In the case of complexation of L2 with any of the metal ions, the
L2/M ratio was retained as 1.0, albeit at a high ligand density. This
trend might be explained by the nature of the carboxylate donor.
The anionic character eliminates the need for external counter
anions to reduce the steric and charge repulsions and the bridging
character preferentially leads to the formation of [(ML2)n] clusters,
according to the molecular structures of the relating homogeneous
metal/L2 complexes mentioned in the section 3.2.2.
3.2. Preparation of homogeneous catalysts
3.2.1. Metal complexes with the L1 ligand
Owing to the difficulty of characterization of metal complex
sites on heterogeneous catalysts, investigation on catalytic reac-
tivity of corresponding homogeneous metal complex system can
be an important reference to estimate the environment of the
active site. In this work, we expanded our study to include man-
ganese(II), iron(II), iron(III), and copper(II) ions in addition to the
previously reported nickel(II) and cobalt(II) complexes [24,25].
In the previous work, [Ni(OAc)(L1)](BPh4), [Ni(L1)2](BF4)2, and
[Co(OAc)(L1)](BPh4) were prepared and their molecular structures
were characterized by X-ray crystallography (Scheme 5). The pre-
pared homogeneous metal complex catalysts have been named as
M(A)/Ln, which stands for the name of the metal ion (M) in the
original metal source, the anions (A) from the metal sources or
ion exchange process, and the name of the ligand (L1 or L2). We
gave priority to the direct comparison of the reactivity between
homogeneous and immobilized catalysts. Therefore, we applied
the complexes directly for the catalytic reaction without further
purification and characterization, except some successful cases.
MnII(Cl)/L1 was prepared from L1 and manganese(II) chlo-
ride in methanol under argon atmosphere. ESI+-MS spectrum of
MnII(Cl)/L1 showed a fragment peak at m/z = 425.97 (Mn + L1 + Cl)+
indicating that the complex consisted of manganese, chlo-
ride, and L1 (Fig. S7). Similar manganese complexes with
the tetradentate tpa ligand such as [MnIICl2(tpa)] [37,38],
[MnIICl(tpa)](ClO4), [(tpa)MnII-(␮-Cl)2-MnII(tpa)](ClO4)2 [39], and
[(tpa)MnII-(␮2-1,3-OAc)2-MnII(tpa)](BPh4)2 (MnII(OAc,BPh4)/tpa)
[8] have been reported. With a tridentate ligand N,N-bis(2-
pyridylmethyl)propargylamine (L3), the molecular structure of
a Cl-bridged hexacoordinate complex with the N3Cl3 donor set,
[(L3)MnII(Cl)-(␮-Cl)2-MnII(Cl)(L3)], has been revealed [40]. Based
on the data from these complexes with similar ligands, MnII(Cl)/L1
may be constructed by hexa-coordinating the N3Cl3 or N4Cl2 ligand
set with or without the coordination of the triazole side arm.
The iron(II) complex FeII(OTf)/L1 was prepared from L1
and iron(II) triflate in acetonitrile under argon atmosphere.
ESI+-MS spectrum of FeII(OTf)/L1 showed a fragment peak at
m/z = 541.10 (Fe + L1 + OTf)+ indicating that the complex con-
sisted of iron, triflate, and L1 (Fig. S8). An iron(III) complex,
We evaluated the ratio of the ligand to metal ions (Ln/M) in order
to estimate the structure of the metal complexes in the immobi-
lized catalyst. Generally, the limitation of the metal loadings on the
site-dense surface can be related to the steric hindrance or charge
repulsion. Furthermore, the metal loadings should be related to the
structural variation of the metal complexes, such as the formation
of the anion bridged cluster [L-M-(␮-anion)-M-L] under the ligand
dense situation [8].
In the low ligand density silica, M(A)/SBA*-Ln-0.5, the Ln/M
values were close to 1.0 for all metal ions studied in this work,
indicating that one isolated ligand site may associate with one
metal ion to form the [M1(Ln)1] species. In the high ligand den-
sity silica, M(A)/SBA*-Ln-4, the Ln/M values were greater than 1.0

J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
19
amine, in addition to the two chloride anions to form octahedral
geometry (Fig. 1b). The pyridine moiety locates the trans posi-
tion of the other pyridine. The molecular structures of related iron
complexes, namely, [FeII(OTf)(tpa)], [FeIIICl3(L1Bz)], [FeIIICl3(L3)],
[FeIIICl2(L1Bz)](PF6), and [(L3)FeIII(Cl)-(␮-Cl)2-FeIII(Cl)(L3)] have
also been reported, where L1Bz is the benzyl substituted analogue
of L1 at the tBu group on the triazole ring and L3 is N,N-bis(2-
pyridylmethyl)propargylamine [41–43]. The complexes with the
tridentate ligand L3 require further coordination of the third chlo-
ride anion to form a N3Cl3 donor set, as shown in Fig. 1a. In contrast,
the L1 and L1Bz ligands also form complexes with a N4Cl2 donor set,
as shown in Fig. 1b, with the tetradentate coordination of the lig-
and after the anion exchange treatment to remove one chloride
anion. This structural change represents the low donor ability of
the triazole nitrogen.
A copper(II) complex, CuII(OAc,BF4)/L1 ([CuII(OAc)(L1)](BF4)),
was prepared from L1 and copper(II) acetate in methanol with
subsequent anion exchange with NaBF4. ESI+-MS spectrum of the
copper complex showed fragment peaks at m/z = 399.12 (Cu + L1)+
and 458.12 (Cu + L1 + OAc)+ indicating that the complex consisted
of copper, acetate, and the L1 ligand (Fig. S11). The molecular struc-
ture of the complex was revealed by X-ray crystallography (Fig. 2).
It was found that the crystal consists of two units and both units
independently form one-dimensional chain structures by the coor-
dination of the triazole group to the next Cu ion. In both units,
the copper centers are surrounded by the N4O donor set com-
prised of three nitrogen donors of the ligand, one monodentate
carboxylate, and the triazole donor from another ligand. In contrast
to our complex, intramolecular coordination of the triazole group
was observed in a related complex, [CuIICl(L1Bz)](ClO4) [44]. These
structures indicate that the intra- and intermolecular coordination
of the triazole group in Cu(L1)-type complexes may be converted
by small structural differences.
Scheme 5. Preparation of Ni and Co complexes with L1 in previous works [24].
Until now, several metal complexes with L1-type ligands with
or without coordination of the triazole group have been reported.
These include the [ZnII(MeCN)(L1Ar)] (Ar = aromatic groups on the
triazole ring), [ZnII(L1Ar)2] [36,45,46], fac-[ReI or TcI(CO)3(L1Bz)]
[47], and RuII(L1) complexes [48].
3.2.2. Metal complexes with L2 ligand
In order to prepare M(A)/L2, we applied MnIICl2, FeIIICl3,
CoIICl2, NiII(OAc)2, and CuIICl2 as metal sources. In ESI+-MS
spectra, the MII(A)/L2 complexes showed fragment peaks of
(M + L2 + solvent)+ and the FeIII(Cl)/L2 complex showed fragment
peaks of (Fe + L2 + Cl)+ and (3Fe + 3L2+ 3Cl), indicating that the com-
plex consisted of metal, chloride, and L2 ligand (Figs. S12–S16). The
MS spectra also support the presence of cluster complexes formed
as a result of the bridging nature of the carboxylate moiety of the
L2 ligand. As with the complexes containing L1, we applied the
M(A)/L2 complexes directly for catalysis without further purifica-
tion and characterization to give priority to a direct comparison of
the reactivity between homogeneous and immobilized catalysts.
To date, structural characterizations of several metal com-
plexes with L2-type ligands have been achieved. For examples,
structures of [{MnII(␮-L4)(OH2)2}2](ClO4)2 (L4 = N,N-bis(2-
pyridylmetyl)glycine) [49], [(L4)MnIII-(␮-O)2-MnIV(L4)](ClO4)
[50], [(L4)Fe(OH2)-(␮-O)-Fe(OH2)(L4)](ClO4)2, [(L4)3FeIII3-(␮-
O)2-(␮-OH)](ClO4) [51,52], [FeII3(OAc)3(␮-L4)3](ClO4)[53],
Fig. 1. Molecular structures of (a) [FeIIICl3(L1)] and (b) [FeIIICl2(L1)](BPh4). Hydro-
gen atoms and BPh4 counter anion have been omitted for clarity. Displacement
ellipsoids are drawn at the 30% probability level.
Fe(Cl)/(L1) ([FeIIICl3(L1)]), was prepared from L1 and iron(III) chlo-
ride in methanol. The ESI+-MS spectrum of FeIII(Cl)/L1 showed
fragment peaks at m/z = 458.03 (Fe + L1 + Cl + MeO)+ and 462.03
(Fe + L1 + 2Cl)+, indicating that the complex consisted of iron, chlo-
ride, and L1 ligand (Fig. S9). X-ray crystallography of Fe(Cl)/(L1)
revealed its molecular structure, in which the iron center is
coordinated by the facially oriented three nitrogen donors of
the two pyridine side arms and the tertiary amine, in addi-
tion to the three chloride anions in an octahedral geometry
(Fig. 1a). FeIII(Cl,BPh4)/L1 ([FeIIICl2(L1)](BPh4)) was prepared by
anion exchange of the in situ generated complex FeIII(Cl)/L1 and
NaBPh4 in methanol. The ESI+-MS spectrum of FeIII(Cl,BPh4)/L1
showed a fragment peak at m/z = 458.03 (Fe + L1 + Cl + MeO)+, which
indicated that the complex consisted of iron, chloride, and the
L1 ligand (Fig. S10). The molecular structure of FeIII(Cl,BPh4)/L1
was clarified by X-ray crystallography (Fig. 1b). In this complex,
the iron center is coordinated by the four nitrogen donors of the
two pyridine side arms, the triazole nitrogen, and the tertiary
[(L4)Fe-(␮-O)-(␮-OH)-Fe(L4)](ClO4)2,
{[Cu(␮-L4)](ClO4)}n
[54], [CuIICl(L4)] [55], arboxylate bridged polynuclear Mn, Co,
and Ni complexes of L5 (L5 = N-(2-pyridylmethyl)iminodiacetate)
[56], Cs2[Fe2(L5)2(O)(CO3)] [57], polynuclear cobalt complex
of L6 (L6 = N-(2-pyridylmethyl)glycine) [58], [Cu2Cl2(L6)2] [59],
[Cu(L6)2] [60], [Ni(L6)2] [61], [ReI(CO)3(L7)] (L7 = N-(propargyl)-
N-(2-pyridylmethyl)glycine), and [ReI(CO)3(L2Bz)] (L2Bz = benzyl
substituted L2 ligand at the tBu position) [62] have been revealed.

20
J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
Scheme 6. Cyclohexane oxidation.
the metal-acylperoxo complex. The alcohol selectivity is calculated
as [A]/([K] + [L]). The catalyst turn over number [TON = (product;
␮mol)/(metal; ␮mol)] is calculated from the amounts of the prod-
uct determined by GC and the metal loadings obtained by atomic
absorption spectroscopy [8].
3.3.1. Reactivity of homogeneous complex catalysts
The catalytic activities of homogeneous metal complexes
and corresponding metal sources for cyclohexane oxida-
tion with mCPBA as the oxidant are shown in Table 3. The
courses of the formation of the product with time using the
homogeneous catalysts are shown in Figs. S17 and S18 . The
various homogeneous metal complex catalysts used in this
work are listed in Table
1 according to the nomenclature,
M(A)/Ln. In addition, the results of previously reported com-
plex catalysts, NiII(OAc)/L1 {=[Ni(OAc)2(H2O)(L1)]}, NiII(BF4)/L1
{=[Ni(L1)2](BF4)2}, CoII(OAc,BPh4)/L1 {=[Co(OAc)(L1)](BPh4)},
NiII(OAc,BPh4)/tpa
{=[Ni(OAc)(tpa)(H2O)](BPh4)},
CoII(OAc/BPh4)/tpa {=[Co(OAc)(TPA)](BPh4)}, FeII(OAc,BPh4)/tpa
{=[Fe2(␮-OAc)2(tpa)2](BPh4)}, and MnII(OAc,BPh4)/tpa {=[Mn2(␮-
OAc)2(tpa)2](BPh4)} have also been cited for the sake of comparison
[8,24,25]. Under the conditions employed in this work, we con-
firmed that the reaction does not take place with mCPBA in the
absence of the catalysts (entry 26).
Initially, we investigated the difference between the L1 and tpa
ligands for each metal. In the complexes with the L1 ligand, the TON
as well as the selectivity for the alcohol based on the metal ion were
found to follow the order Ni ≈ Co > Fe > Mn > Cu. This trend is similar
to those of the corresponding MII/tpa complexes (entries 1–6 versus
13–16), with the exception of the alcohol selectivity of the CoII
complexes (Ni/L1 > Co/L1 versus Ni/tpa < Co/tpa). The activity trend
indicates that the reaction mechanisms of both MII/L1 and MII/tpa
systems for each metal might be similar [8,63–66]. In the case of
manganese salts and complexes, MnIICl2, MnII(OAc,BPh4)/tpa, and
MnII(Cl)/L1 catalysts showed similar activity, as evaluated by TON,
selectivity of the alcohol, and the reaction rate (entries 1, 13, 17).
In other words, there was no clear effect of the nature of the lig-
and in the case of manganese-based catalytic systems. It was also
Fig. 2. Molecular structure of [CuII(OAc)(L1)](BF4). The crystal contains two inde-
pendent units and both these units separately form one-dimensional chain
structures. Hydrogen atoms, Et2O, and the BF4 counter anion have been omitted
for clarity. Displacement ellipsoids are drawn at the 30% probability level.
In these complexes, the bridging mode of the carboxylate ligand
encourages the variation of the structure of the complexes from
mono- to polynuclear and this nature often makes difficult to
characterize these structures.
3.3. Catalytic oxidation of cyclohexane
In 2007, Itoh and coworkers studied the oxidation of alka-
nes with mCPBA oxidant using the MII/tpa (M = Mn, Fe, Ni, Co)
homogeneous complex catalysts [8]. Based on this report, we
have expanded the study to the immobilization of the nickel
and cobalt complexes with a tpa-type ligand L1 to construct
NiII(OAc)/SBA*-L1-x and CoII(OAc)/SBA*-L1-x heterogeneous cat-
alysts [24,25]. We investigated the catalytic oxidation activity of
the immobilized and homogeneous catalysts with the use of Mn,
Fe Co, Ni, and Cu ions with the L1 and L2 ligands. The reaction
conditions are given in Scheme 6. Typically, cyclohexanol (A), cyclo-
hexanone (K), chlorocyclohexane, and ε-caprolactone (L), which is
the Baeyer-Villiger oxidation product of K, are formed as the oxi-
dation products. In addition, chlorobenzene (C) was formed as a
byproduct from mCPBA by the radical-type O O bond scission of
observed that the initial valence of the iron center (FeII versus FeIII
:
entries 2 versus 3 and 18 versus 19) did not affect the catalytic
activity, suggesting that the initial complexes might convert to
same species in the catalytic cycle [67–69]. The TONs decreased
with the order of chelating ability of the ligand (FeII(OAc,BPh4)/tpa
> FeII(OTf)/L1 > FeII(OTf)2), indicating that chelation prevents the
irreversible formation of inactive ␮-oxo complexes or iron oxide
(entries 3, 14, 19). In the case of cobalt-based catalysts, the Co salts
showed a faster reaction with low selectivity toward the alcohol
[25]. When the chelating ability increased from the ligand L1 in the
CoII(OAc,BPh4)/L1 to tpa in CoII(OAc,BPh4)/tpa, the reaction became
slower with higher selectivity for the alcohol (entries 4, 15, 20, 21)
[8,25,70]. The contribution of the free cobalt ion in the reaction of
CoII(OAc,BPh4)/L1 catalyst may induce a difference in the order of

J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
21
Table 3
Cyclohexane oxidation by homogeneous complex catalysts.a
Entry
Cat.
Time (h)
TONb
A
A/(K + L)
K
L
C
1
2
MnII(Cl)/L1
1
1
1
1
2
1
2
1
2
1
3
3
3
3
1
1
1
3
1
1
1
1
0.5
0.5
1
1
1
1
3
122
303
302
533
604
616
647
16
46
93
212
397
454
9.0
54
41
55
101
98
49
56
4.4
6.7
37
15
72
22
2.1
55
62
16
4.1
17
22
6
14
31
20
1.6
2.5
0.9
20
56
2.1
5.2
3.9
5.0
5.4
7.7
8.5
2.7
5.0
2.5
6.1
3.9
8.7
1.3
1.8
6.4
18.8
FeIII(Cl,BPh4)/L1
167
156
432
459
340
377
24
41
28
106
210
240
16
3
FeII(OTf)/L1
Co(OAc,BPh4)/L1
4c,d
5c
6c
7
Ni(OAc,BPh4)/L1
Ni(BF4)/L1
Cu(OAc,BF4)/L1
8
9
Mn(Cl)/L2
Fe(Cl)/L2
Co(Cl)/L2
Ni(OAc)/L2
10
11
12
13c
14c
15c
29
30
Cu(Cl)/L2
5.0
trace
trace
trace
trace
trace
12
16
5.6
4
f
Mn(OAc,BPh4)/tpa
Fe(OAc,BPh4)/tpa
Co(OAc,BPh4)/tpa
99
—
—
—
—
f
396
300
∼600
587
155
190
205
515
368
1.5
f
f
f
f
—
—
16c
17
18
Ni(OAc,BPh4)/tpa
MnIICl2
69
56
36
24
221
80
0.5
0.5
6.6
1.3
5
—
8.5
2.3
3.7
6.9
2.3
4.2
2.5
2.0
1.0
1.1
1.6
f
67
FeIIICl3
111
222
511
277
15
10
16
8
15
19
FeII(OTf)2
CoIICl2
20c,d
21c,d
22
CoII(OAc)2
NiIICl2
7
0.1
0.1
2.9
1.8
0
23
24
NiII(OAc)2
CuIICl2
1.2
9.1
3.3
8
25
CuII(OAc)2
None
26c,d,e
a
General conditions: Catalyst (2 ␮mol), cyclohexane (15 mmol), mCPBA (2 mmol), CH2Cl2 (3 mL), MeCN (1 mL) at room temperature under Ar.
b
c
d
e
f
TON = [product]/[metal].
Obtained from references [8,24,25].
At 35 ◦C.
Assumed TON.
No data.
the alcohol selectivity between MII/L1 and MII/tpa (Ni/L1 > Co/L1
versus Ni/tpa < Co/tpa). The L1 and tpa nickel complexes showed
high TONs and product selectivity while almost no activity of its
simple salts was observed at room temperature (entries 5, 6, 16,
22, 23) [24]. Coordination of a multidentate ligand, especially with
nitrogen donors, generally enhances the catalytic oxidation activity
of nickel system, as described in literature [14,15,71–73].
In the case of copper-based catalysts, the nature of the metal
sources and the L1 complex showed no reactivity (entries 7, 24,
25).
In terms of the metal complexes with the L2 ligand, the order
of TON (entries 8–12; Ni ≈ Co > Fe > Mn) was similar to those of
the L1 catalysts, although the TONs of the M/L2 catalysts were
slightly lower (entries 1–7). In contrast, the trends of the alcohol
selectivity were different, depending on the nature of the ligand
(Ni > Co ≈ Fe > Mn for L1; Ni > Fe > Co > Mn for L2; Co > Ni > Fe > Mn
for tpa). These differences in catalytic activity depending on the
nature of the ligand can also be influenced by the coordination
structure of the metal complexes, although this effect is unclear.
Nevertheless, this data for homogeneous catalysts can be useful for
the evaluation of the immobilized catalysts.
catalysis was performed by using a fixed amount of the metal ion
(2 ␮mol) in order to facilitate a direct comparison with the homoge-
neous complexes based on the TON values. The catalyses performed
with the simple metal salts were not affected by the addition of
the SBA-N3-0.5 mesoporous silica support (entries 17–25 versus
entries 27–35 in Tables 3 and 4, Co > Fe > Mn » Ni ∼ Cu). In case of
the copper-based catalysts, both homogeneous and heterogeneous
catalysts showed no activity.
The effect of immobilization and ligand density of the cata-
lysts in each metal ion was evaluated by comparison with the
homogeneous analogues. In the L1 system, the TON as well as the
selectivity toward the alcohol in case of the ligand isolated cata-
lysts, M(A)/SBA*-L1-0.5, were dependent on the nature of the metal
ions (Ni ≈ Co > Fe > Mn > Cu) and the observed order was similar to
that for corresponding homogeneous M(A)/L1 system (entries 1, 3,
4, 5, and 7 versus 36, 38, 40, 42, and 44). This similarity suggests
that the ligand isolated environment on the silica support enables
formation of mononuclear active sites similar to the homogeneous
system in addition to the smooth access of the oxidant and sub-
strates to the active site. In contrast, the ligand dense catalysts,
M(A)/SBA*-L1-4, showed a lower TON of the alcohol compared
to the ligand isolated catalysts, except in the case of cobalt cata-
lysts (entries 36–45). In our previous research on Ni/SBA*-L1 [23]
and relating immobilized materials [24], rapid diffusion of reagents
in the mesopore at higher ligand density was confirmed by N2
adsorption experiment as well as extraction of copper ion by a
chelating reagent during the preparation. The ligand dense envi-
ronment induces structural changes such as the formation of inert
[M(L1)2] species or polynuclear cluster complexes, which prevent
the access and coordination of the oxidant and substrates to the
3.3.2. Reactivity of the immobilized catalysts
The results of the cyclohexane oxidation reaction with immo-
bilized M(A)/SBA*-Ln-x catalysts, including previously reported Ni
and Co catalysts, are summarized in Table 4. The loading amounts
of the metal ions on the catalysts were determined before and after
catalysis to evaluate leaching of the metal ions (Tables 1 and 2).
In addition, the progress of product formation with time using
the immobilized catalysts is shown in Figs. S19–25. The oxidation

22
J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
Table 4
Cyclohexane oxidation by immobilized catalysts.a
Entry
Catalyst
L/Mratio
TONb
A
A/(K + L)
K
L
C
27c
28c
29c
30c,d
31c,d
32c
33c
34c
35c
36
MnIICl2 + silica
1
1
1
1
1
1
1
1
134
218
191
422
512
3.0
1.0
4.0
1.1
127
77
396
230
577
745
534
251
18
42
45
19
85
109
1.8
0.4
2.5
1.0
53
60
56
35
138
80
17
10
11
2.5
44
53
20
30
114
68
32
39
7.1
17
8.2
14
20
0.9
0.1
0.9
0.5
5.6
4.1
49
46
6
25
34
10
6.8
1.4
4.0
0.8
39
33
38
39
37
16
38
2.7
3.5
7.0
4.3
4.0
1.1
2.0
1.2
0.7
2.2
1.2
3.8
2.8
4.0
7.1
10.5
12.6
1.0
4.9
1.4
2.0
6.8
3.3
3.3
3.3
4.6
3.9
FeIIICl3 + silica
130
104
199
311
6.9
7.5
12
9.4
58
53
288
151
377
582
247
117
22
26
52
39
202
145
407
198
211
177
FeII(OTf)2 + silica
CoIICl2 + silica
CoII(OAc)2 + silica
NiIICl2 + silica
NiII(OAc)2 + silica
CuIICl2 + silica
CuII(OAc)2 + silica
Mn(Cl)/SBA*-L1-0.5
Mn(Cl)/SBA*-L1-4
Fe(OTf)/SBA*-L1-0.5
Fe(OTf)/SBA*-L1-4
Co(OAc)/SBA*-L1-0.5
Co(OAc)/SBA*-L1-4
Ni(OAc)/SBA*-L1-0.5
Ni(OAc)/SBA*-L1-4
Cu(OAc)/SBA*-L1-0.5
Cu(OAc)/SBA*-L1-4
Mn(Cl)/SBA*-L2-0.5
Mn(Cl)/SBA*-L2-4
Fe(OTf)/SBA*-L2-0.5
Fe(OTf)/SBA*-L2-4
Co(OAc)/SBA*-L2-0.5
Co(OAc)/SBA*-L2-4
Ni(OAc)/SBA*-L2-0.5
Ni(OAc)/SBA*-L2-4
1
1.2
1.5
1.1
1.5
1.0
1.4
0.9
1.7
1.0
1.0
1.0
1.1
1.0
1.3
0.7
1.1
0.9
1.3
37
38
39
40d,e
41d,e
42e
43e
44
45
46
47
48
19
67
106
401
205
497
358
314
215
49
50 d
51 d
52 f
53 f
a
General conditions: Catalyst (2 ␮mol metal), cyclohexane (15 mmol), mCPBA (2 mmol), CH2Cl2 (3 mL), MeCN (1 mL) at room temperature under Ar for 3 h.
b
c
d
e
f
TON = [product]/[M].
With 50 mg of SBA-N3-0.5 silica.
At 35 ◦C.
Obtained from reference [24,25].
Reaction time = 5 h.
active site. In the L2 system, metal ion dependency of the activity
of M(A)/SBA*-L2-0.5 (TON: Co > Fe > Ni > Mn, selectivity of the alco-
hol: Fe > Ni > Co > Mn) was changed from that of the homogeneous
complexes (TON: Ni ≈ Co > Fe > Mn and selectivity of the alcohol:
Ni > Fe > Co > Mn) upon immobilization. This result represents the
difference of the structure of the active species. The ligand isolated
environment may force the formation of the mononuclear species,
even though the bridging character of the carboxylate moiety of L2
ligand preferentially leads to the formation of cluster complexes in
homogeneous solutions.
In manganese-based catalysts, the reaction terminated within
the initial ten minutes (Figs. S17 –S19) and no clear effect of the lig-
and density of immobilized system as well as the difference of the
ligand (L1, L2, or tpa) was observed with any homogeneous, immo-
bilized, and simple salt catalysts (entries 1, 8, 13, 17, 36, 37, 46,
and 47). All of the immobilized manganese catalysts showed severe
leaching of the manganese ion during the reaction (Tables 1 and 2).
We presumed that the observed reaction is mainly catalyzed by the
leached manganese ion in the presence of excess mCPBA.
In case of the iron catalysts, the reactive mononuclear site of
L2 may be selectively formed on the site-isolated support, even
though L2 prefers to form cluster complexes even in its homo-
geneous solution. The TONs of the iron catalysts for production
of alcohols changed in the following order: FeII(OTf)/SBA*-L2-0.5
(TON = 401) ≈ FeII(OTf)/SBA*-L1-0.5
(TON = 302) > FeII(OTf)/SBA*-L1-4
(TON = 212) ≈ FeII(OTf)/SBA*-L2-4
(TON = 396) > FeII(OTf)/L1
(TON = 230) ≈ FeII(Cl)/L2
(TON = 205) ≈ FeII(OTf)2
Fig. 3. Plausible structure of the iron L2 complex sites on the surface.
especially for the L2 ligand. These results suggest that the isolated
immobilization of L1 and L2 prevents the formation of inactive
species such as anion-bridged (e.g. ␮-oxo, ␮-acetato) clusters and
[Fe(L)2] (Fig. 3).
(TON = 205) (entries 3, 9, 19, 38, 39, 48, and 49). Furthermore,
leaching of the iron ions was suppressed in the immobilized L2
systems, in contrast with the L1 systems (Tables 1 and 2). In other
words, the ligand isolated sites showed higher reactivity than the
ligand dense sites, homogeneous complexes, and free metal ion,
In the previous research on cobalt L1 system [25], the
ligand dense catalyst, CoII(OAc)/SBA*-L1-4, showed higher selec-

J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
23
tivity of the alcohol compared to the ligand isolated catalyst,
Acknowledgments
CoII(OAc)/SBA*-L1-0.5, and cobalt salts (entries 31, 40, and 41). We
considered that suppression of the cobalt leaching on the ligand
dense environment resulted in expression of the selective catalysis
on the Co(L1) site. Homogeneous and heterogeneous Co(L2) sys-
tems showed relatively similar TON and the selectivity toward the
alcohol as those for cobalt salts with the suppression of the reac-
tion rate. In addition to the observation of the leaching of cobalt
ion, Co(L2) sites were less reactive than the Co(L1) sites.
This work was supported in part by the Cooperative Research
Program of the “Network Joint Research Center for Materials and
Devices” [Grant number 2012348] and by the Strategic Devel-
opment of Research Infrastructure for Private Universities [Grant
number S1201017] grant from MEXT, Japan. We also gratefully
acknowledge the support from JSPS KAENHI [Grants JP16K17904
and JP26420788] and CREST, JST [Grant number JPMJCR16P1]. J. N.
and S. H. also thank Kanagawa University for financial support.
The activity of immobilized Ni catalysts, NiII(OAc)/SBA*-L1-
x, has been previously reported to decrease upon increase in
the surface ligand density (x = 0.5 to 4 mol%). We proposed that
the catalyst lost the activity of the isolated site upon increase
in the coordinatively inert and saturated [Ni(L1)2] site. In both
homogeneous and heterogeneous systems, the Ni(L2) catalysts
showed a relatively lower TON and reaction rate compared to
corresponding L1 catalysts. In terms of the selectivity of the
alcohol in the immobilized systems, NiII(OAc)/SBA*-L2-x catalysts
showed a lower selectivity for the alcohol (A/(K + L) = 4–5) than
NiII(OAc)/SBA*-L1-x catalysts (A/(K + L) = 10–13). This relation is
opposite to that observed for homogeneous analogues, NiII(OAc)/L2
(A/(K + L) = 8.7) > NiII(OAc)/L1 (A/(K + L) = 7.7), indicating that the
coordination mode of these immobilized complexes can be differ-
ent from the homogeneous analogues.
We did not carry out the reuse tests of the immobilized cata-
lysts because partial leaching of the metal ions was observed in
all of the immobilized complexes of L1 and L2 after catalysis. The
metal leaching occurs either because of the oxidation of the immo-
bilized ligands or as a result of the abstraction of the metal ions
with coordinating molecules such as mCPBA, m-chlorobenzoate,
and cyclohexanol.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.mcat.2017.09.
027.
References
[1] H. Hussain, A. Al-Harrasi, I.R. Green, I. Ahmed, G. Abbas, N.U. Rehman, RSC
Adv. 4 (2014) 12882–12917.
[2] W. Nam, J.Y. Ryu, I. Kim, C. Kim, Tetrahedron Lett. 43 (2002) 5487–5490.
[3] T. Newhouse, P.S. Baran, Angew. Chem. Int. Ed. 50 (2011) 3362–3374.
[4] H. Lu, X.P. Zhang, Chem. Soc. Rev. 40 (2011) 1899–1909.
[5] G. Shul’pin, D. Loginov, L. Shul’pina, N. Ikonnikov, V. Idrisov, M. Vinogradov, S.
Osipov, Y. Nelyubina, P. Tyubaeva, Molecules 21 (2016) 1593.
[6] M. Bordeaux, A. Galarneau, J. Drone, Angew. Chem. Int. Ed. 51 (2012)
10712–10723.
[7] A.E. Shilov, G.B. Shul’pin, Chem. Rev. 97 (1997) 2879–2932.
[8] T. Nagataki, Y. Tachi, S. Itoh, Chem. Commun. (2006) 4016–4018.
[9] T. Nagataki, K. Ishii, Y. Tachi, S. Itoh, Dalton Trans. (2007) 1120–1128.
[10] T. Nagataki, S. Itoh, Chem. Lett. 36 (2007) 748–749.
[11] F.F. Pfaff, F. Heims, S. Kundu, S. Mebs, K. Ray, Chem. Commun. 48 (2012)
3730–3732.
[12] S. Hikichi, K. Hanaue, T. Fujimura, H. Okuda, J. Nakazawa, Y. Ohzu, C.
Kobayashi, M. Akita, Dalton Trans. 42 (2013) 3346–3356.
[13] J. Nakazawa, S. Terada, M. Yamada, S. Hikichi, J. Am. Chem. Soc. 135 (2013)
6010–6013.
[14] M. Balamurugan, R. Mayilmurugan, E. Suresh, M. Palaniandavar, Dalton Trans.
40 (2011) 9413–9424.
4. Conclusions
In this work, homogeneous and immobilized metal complex
catalysts with two (pyridylmethyl)amine type ligands were pre-
pared and their catalytic activity for cyclohexane oxidation with
mCPBA oxidant was evaluated. In homogeneous complexes of dif-
ferent metal ions with ligand L1, the decreasing order of TONs
and selectivity toward cyclohexanol was Ni ≈ Co > Fe > Mn > Cu. This
trend was essentially similar to that observed for correspond-
ing M/tpa complexes. The metal complexes with the L2 ligand
showed a similar order of TONs for the formation of cyclohex-
anol (Ni ≈ Co > Fe > Mn) to those of the L1 catalysts although the
TON values for M/L2 were slightly lower. In contrast, the trends
of the alcohol selectivity with complexes of L1, L2, and tpa lig-
ands were different, (Ni > Co ≈ Fe > Mn for L1, Ni > Fe > Co > Mn for
L2, and Co > Ni > Fe > Mn for tpa). These differences in reactivity may
be attributed to the coordination structure of the metal complexes,
although these coordination structures are as yet unclear. The reac-
tivity of the immobilized complex catalysts depends on the metal
ion, ligand, and the ligand density. In the case of non-aggregated
catalysts such as the L1 complexes, the site-isolated condition is
required to transfer the homogeneous catalyst to the immobilized
system without change in its activity. On the other hand, the lig-
and dense immobilized condition generates aggregated sites that
are never formed in the homogeneous solution. In the L2 system,
the metal ion dependency of the reactivity of M(A)/SBA*-L2-0.5
changed when the homogeneous complexes were immobilized.
The isolated environment of the L2 ligand may force the formation
of the mononuclear complex species, although the bridging charac-
ter of the carboxylate moiety of L2 ligand preferentially leads to the
formation of cluster complexes. Finally, this result clearly demon-
strates the importance of the density control of the surface sites for
constructing the desired catalyst on a solid support.
[15] M. Sankaralingam, M. Balamurugan, M. Palaniandavar, P. Vadivelu, C.H.
Suresh, Chem. Eur. J. 20 (2014) 11346–11361.
[16] E. Tordin, M. List, U. Monkowius, S. Schindler, G. Knör, Inorg. Chim. Acta 402
(2013) 90–96.
[17] T.J. Terry, G. Dubois, A. Murphy, T.D.P. Stack, Angew. Chem. Int. Ed. 46 (2007)
945–947.
[18] J. Hong, K.E. Djernes, I. Lee, R.J. Hooley, F. Zaera, ACS Catal. 3 (2013)
2154–2157.
[19] Y. Feng, E.G. Moschetta, C.W. Jones, Chem. Asian J. 9 (2014) 3142–3152.
[20] Y. Aratani, Y. Yamada, S. Fukuzumi, Chem. Commun. 51 (2015) 4662–4665.
[21] M. Yamada, K.D. Karlin, S. Fukuzumi, Chem. Sci. 7 (2016) 2856–2863.
[22] J. Nakazawa, T.D.P. Stack, J. Am. Chem. Soc. 130 (2008) 14360–14361.
[23] J. Nakazawa, B.J. Smith, T.D.P. Stack, J. Am. Chem. Soc. 134 (2012) 2750–2759.
[24] J. Nakazawa, T. Hori, T.D.P. Stack, S. Hikichi, Chem. Asian J. 8 (2013)
1191–1199.
[25] J. Nakazawa, A. Yata, T. Hori, T.D.P. Stack, Y. Naruta, S. Hikichi, Chem. Lett. 42
(2013) 1197–1199.
[26] V. Malgras, Q. Ji, Y. Kamachi, T. Mori, F.-K. Shieh, K.C.-W. Wu, K. Ariga, Y.
Yamauchi, Bull. Chem. Soc. Jpn. 88 (2015) 1171–1200.
[27] E. Yamamoto, K. Kuroda, Bull. Chem. Soc. Jpn. 89 (2016) 501–539.
[28] K.S. Lakhi, D.-H. Park, K. Al-Bahily, W. Cha, B. Viswanathan, J.-H. Choy, A. Vinu,
Chem. Soc. Rev. 46 (2017) 72–101.
[29] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D.
Stucky, Science 279 (1998) 548–552.
[30] T.L. Mindt, H. Struthers, B. Spingler, L. Brans, D. Tourwé, E. García-Garayoa, R.
Schibli, ChemMedChem 5 (2010) 2026–2038.
[31] C.H. Zhang, X.D. Shen, R. Sakai, M. Gottschaldt, U.S. Schubert, S. Hirohara, M.
Tanihara, S. Yano, M. Obata, N.O. Xiao, T. Satoh, T. Kakuchi, J. Polym. Sci. Part
A: Polym. Chem. 49 (2011) 746–753.
[32] M. Costas, M.P. Mehn, M.P. Jensen, L. Que, Chem. Rev. 104 (2004) 939–986.
[33] E.Y. Tshuva, S.J. Lippard, Chem. Rev. 104 (2004) 987–1012.
[34] M.M. Abu-Omar, A. Loaiza, N. Hontzeas, Chem. Rev. 105 (2005) 2227–2252.
[35] S.V. Kryatov, E.V. Rybak-Akimova, S. Schindler, Chem. Rev. 105 (2005)
2175–2226.
[36] J.T. Simmons, J.R. Allen, D.R. Morris, R.J. Clark, C.W. Levenson, M.W. Davidson,
L. Zhu, Inorg. Chem. 52 (2013) 5838–5850.
[37] Y. Hitomi, A. Ando, H. Matsui, T. Ito, T. Tanaka, S. Ogo, T. Funabiki, Inorg.
Chem. 44 (2005) 3473–3478.
[38] C. Duboc, T. Phoeung, S. Zein, J. Pécaut, M.-N. Collomb, F. Neese, Inorg. Chem.
46 (2007) 4905–4916.

24
J. Nakazawa et al. / Molecular Catalysis 443 (2017) 14–24
[39] B.K. Shin, M. Kim, J. Han, Polyhedron 29 (2010) 2560–2568.
[40] J. Chaignon, S.-E. Stiriba, F. Lloret, C. Yuste, G. Pilet, L. Bonneviot, B. Albela, I.
Castro, Dalton Trans. 43 (2014) 9704–9713.
[41] A.L. Ward, L. Elbaz, J.B. Kerr, J. Arnold, Inorg. Chem. 51 (2012) 4694–4706.
[42] B. Xu, W. Zhong, Z. Wei, H. Wang, J. Liu, L. Wu, Y. Feng, X. Liu, Dalton Trans. 43
(2014) 15337–15345.
[58] T. Ama, T. Yonemura, M. Yamaguchi, Bull. Chem. Soc. Jpn. 79 (2006)
1063–1065.
[59] B. Zheng, H. Liu, J. Feng, J. Zhang, Appl. Organomet. Chem. 28 (2014) 372–378.
[60] X. Wang, J.J. Vittal, Inorg. Chem. 42 (2003) 5135–5142.
[61] Q.-Q. Zhang, Z.-H. Zhang, B.-H. Qu, Q. Chen, M.-Y. He, Inorg. Chim. Acta 418
(2014) 59–65.
[43] E. Gu, W. Zhong, H. Ma, B. Xu, H. Wang, X. Liu, Inorg. Chim. Acta 444 (2016)
159–165.
[62] S.C. Bottorff, A.L. Moore, A.R. Wemple, D.-K. Bucˇar, L.R. MacGillivray, P.D.
Benny, Inorg. Chem. 52 (2013) 2939–2950.
[44] W.S. Brotherton, H.A. Michaels, J.T. Simmons, R.J. Clark, N.S. Dalal, L. Zhu, Org.
Lett. 11 (2009) 4954–4957.
[45] H.A. Michaels, C.S. Murphy, R.J. Clark, M.W. Davidson, L. Zhu, Inorg. Chem. 49
(2010) 4278–4287.
[46] S. Huang, R.J. Clark, L. Zhu, Org. Lett. 9 (2007) 4999–5002.
[47] A.L. Moore, D.-K. Bucar, L.R. MacGillivray, P.D. Benny, Dalton Trans. 39 (2010)
1926–1928.
[48] F. Weisser, H. Stevens, J. Klein, M. van der Meer, S. Hohloch, B. Sarkar, Chem.
Eur. J. 21 (2015) 8926–8938.
[49] H. Iikura, T. Nagata, Inorg. Chem. 37 (1998) 4702–4711.
[50] M. Suzuki, H. Senda, Y. Kobayashi, H. Oshio, A. Uehara, Chem. Lett. 17 (1988)
1763–1766.
[51] H. Furutachi, Y. Ohyama, Y. Tsuchiya, K. Hashimoto, S. Fujinami, A. Uehara, M.
Suzuki, Y. Maeda, Chem. Lett. 29 (2000) 1132–1133.
[52] M.N. Mortensen, B. Jensen, A. Hazell, A.D. Bond, C.J. McKenzie, Dalton Trans.
(2004) 3396–3402.
[63] M.H. Lim, J.-U. Rohde, A. Stubna, M.R. Bukowski, M. Costas, R.Y.N. Ho, E.
Münck, W. Nam, L. Que, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 3665–3670.
[64] M. Costas, K. Chen, L. Que Jr., Coord. Chem. Rev. 200–202 (2000) 517–544.
[65] K.P. Bryliakov, E.P. Talsi, Coord. Chem. Rev. 276 (2014) 73–96.
[66] M. Mitra, H. Nimir, D.A. Hrovat, A.A. Shteinman, M.G. Richmond, M. Costas, E.
Nordlander, J. Mol. Catal. A: Chem. 426 Part B (2017) 350–356.
[67] N.Y. Oh, M.S. Seo, M.H. Lim, M.B. Consugar, M.J. Park, J.-U. Rohde, J. Han, K.M.
Kim, J. Kim, J.L. Que, W. Nam, Chem. Commun. (2005) 5644–5646.
[68] K. Ray, S.M. Lee, L. Que Jr., Inorg. Chim. Acta 361 (2008) 1066–1069.
[69] B. Wang, Y.-M. Lee, M. Clémancey, M.S. Seo, R. Sarangi, J.-M. Latour, W. Nam, J.
Am. Chem. Soc. 138 (2016) 2426–2436.
[70] R.R. Reinig, D. Mukherjee, Z.B. Weinstein, W. Xie, T. Albright, B. Baird, T.S.
Gray, A. Ellern, G.J. Miller, A.H. Winter, S.L. Bud’ko, A.D. Sadow, Eur. J. Inorg.
Chem. 2016 (2016) 2486–2494.
[71] D.D. Narulkar, A.R. Patil, C.C. Naik, S.N. Dhuri, Inorg. Chim. Acta 427 (2015)
248–258.
[53] S.K. Mandal, V.G. Young, L. Que, Inorg. Chem. 39 (2000) 1831–1833.
[54] K.-Y. Choi, Y.-M. Jeon, H. Ryu, J.-J. Oh, H.-H. Lim, M.-W. Kim, Polyhedron 23
(2004) 903–911.
[55] T. Okuno, S. Ohba, Y. Nishida, Polyhedron 16 (1997) 3765–3774.
[56] Z. Chen, Y. Li, C. Jiang, F. Liang, Y. Song, Dalton Trans. (2009) 5290–5299.
[57] Y. Nishida, A. Goto, T. Akamatsu, S. Ohba, T. Fujita, T. Tokii, S. Okada, Chem.
Lett. 23 (1994) 641–644.
[72] M. Sankaralingam, P. Vadivelu, E. Suresh, M. Palaniandavar, Inorg. Chim. Acta
407 (2013) 98–107.
[73] T. Corona, F.F. Pfaff, F. Acun˜a-Parés, A. Draksharapu, C.J. Whiteoak, V.
Martin-Diaconescu, J. Lloret-Fillol, W.R. Browne, K. Ray, A. Company, Chem.
Eur. J. 21 (2015) 15029–15038.