Using polyethyleneimine (PEI) as a scaffold to construct mimicking systems of [FeFe]-hydrogenase: Preparation, characterization of PEI-based materials, and their catalysis on proton reduction

Article abstract of DOI:10.1002/aoc.2979

Ten branched polymeric materials (PEI-P-Fe₂s) derived from polyethyleneimine (PEI) functionalized with [Fe₂(CO) ₅]-units to mimic [FeFe]-hydrogenase were prepared. Before the functionalization, PEI was first premodified using diphenylphosphinamine (NPPh₂) group. In the premodification, three approaches were employed: (i) using PEI with an average molecular weight of 1800 and 600, respectively; (ii) grafting NPPh₂ group by either direct reaction of chlorodiphenylphosphine with PEI or Br(CH₂)₁₁OPPh ₂; and (iii) further premodification with BrCH₂COOH after immobilization of the NPPh₂ group. Reaction of the premodified PEI with diiron hexacarbonyl complexes, [Fe₂(μ-S)₂(CO) ₆] (1), or [Fe₂(μ-S₂C₂H ₄)(CO)₆] (3) produced 10 functionalized materials, PEI-P-Fe₂s. These materials were characterized using a variety of spectroscopic techniques, FTIR, NMR, TGA, and cyclic voltammetry. Spectral comparison with two control complexes, [Fe₂(μ-S) ₂(CO)₅PPh₃] (2) and [Fe₂(μ-S ₂C₂H₄)(CO)₅PPh₃] (4), suggested that the immobilized diiron units of PEI-P-Fe₂s were dominantly pentacarbonyl analogous to complexes 2 and 4, although tetracarbonyl units may also exist because the amine groups of PEI could also be involved in substituting CO, as was the NPPh₂ group. The catalysis of these materials on proton reduction was examined in 0.1 mol l^-1 [NBut ₄]BF₄/DMF containing acetic acid by using cyclic voltammetry. Our results indicated that both the presence of carboxylic acid and dangling the diiron units at the end of a long aliphatic chain improved catalytic efficiency by one-fold. The improvement was attributed to the increase in flexibility of the catalytic center and enhancement of proton transfer during the catalysis. Copyright 2013 John Wiley & Sons, Ltd. Functionalised polyethyleneimine was employed as a scaffold to support diiron mimics. The amine group of PEI and the carboxylic acid grafted onto its skeleton could serve as a proton relay in catalysis on proton reduction and hence improve catalytic efficiency. Copyright

Full text of DOI:10.1002/aoc.2979

Full Paper
Received: 18 December 2012
Revised: 22 January 2013
Accepted: 26 January 2013
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/aoc.2979
Using polyethyleneimine (PEI) as a scaffold to
construct mimicking systems of [FeFe]-
hydrogenase: preparation, characterization of
PEI-based materials, and their catalysis on
proton reduction
Yao Lia, Wei Zhongb, Guifen Qiana, Zhiyin Xiaob and Xiaoming Liua,b*
Ten branched polymeric materials (PEI-P-Fe2s) derived from polyethyleneimine (PEI) functionalized with [Fe2(CO)5]-units to mimic
[FeFe]-hydrogenase were prepared. Before the functionalization, PEI was first premodified using diphenylphosphinamine (NPPh2)
group. In the premodification, three approaches were employed: (i) using PEI with an average molecular weight of 1800 and 600,
respectively; (ii) grafting NPPh2 group by either direct reaction of chlorodiphenylphosphine with PEI or Br(CH2)11OPPh2; and (iii)
further premodification with BrCH2COOH after immobilization of the NPPh2 group. Reaction of the premodified PEI with diiron
hexacarbonyl complexes, [Fe2(m-S)2(CO)6] (1), or [Fe2(m-S2C2H4)(CO)6] (3) produced 10 functionalized materials, PEI-P-Fe2s. These
materials were characterized using a variety of spectroscopic techniques, FTIR, NMR, TGA, and cyclic voltammetry. Spectral compar-
ison with two control complexes, [Fe2(m-S)2(CO)5PPh3] (2) and [Fe2(m-S2C2H4)(CO)5PPh3] (4), suggested that the immobilized diiron
units of PEI-P-Fe2s were dominantly pentacarbonyl analogous to complexes 2 and 4, although tetracarbonyl units may also exist
because the amine groups of PEI could also be involved in substituting CO, as was the NPPh2 group. The catalysis of these materials
on proton reduction was examined in 0.1 mol lꢀ1 [NBut4]BF4/DMF containing acetic acid by using cyclic voltammetry. Our results in-
dicated that both the presence of carboxylic acid and dangling the diiron units at the end of a long aliphatic chain improved catalytic
efficiency by one-fold. The improvement was attributed to the increase in flexibility of the catalytic center and enhancement of pro-
ton transfer during the catalysis. Copyright © 2013 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this article.
Keywords: [FeFe]-hydrogenase; PEI-based materials; proton reduction; catalysis
In the past decade, a large number of diiron complexes of the
Introduction
core, [Fe2(CO)6ꢀx] (x = 1, 2), have been prepared.[5–13] Those
The structure of [FeFe]-hydrogenase, a metalloenzyme catalyzing
hydrogen evolution reversibly and rapidly, was revealed over
one decade ago. The enzyme exhibits one significant feature,
performing both hydrogen oxidation and evolution by employing
those elements widely available and relatively abundant in the
crust of our planet, e.g. iron, sulfur (Fig. 1).[1–4] This is in striking
contrast to the practice in industry where precious metals –
platinum and palladium – are required. These well-known metals
are in extremely low abundance.
The extraordinary feature of the natural system and the rele-
vance of hydrogen as a future energy vector have tremendously
inspired synthetic chemists to mimic the dimetallic center of the
enzyme. Indeed, exploring bio-inspired mimetic chemistry has
not attempted to duplicate the active site. Instead, the main
concerns of modeling the metal center of the enzyme lie in (i)
understanding the catalytic mechanism of the enzyme at the
molecular level and, more specifically, how this natural ‘machinery’
consisting of ‘cheap building blocks’ can achieve its marvelous
catalytic function, and thus, general principles of the enzymatic
catalysis can be drawn; and (ii) being inspired by those general
principles, artificial systems may be constructed to have functional-
ities comparable to the natural system.
diiron complexes, in one way or another, resemble the diiron
center of the enzyme. In mimicking the natural systems, particu-
lar attention has been paid to the preparation of mimics
possessing a basic group either directly coordinating to the
diiron center or locating in the secondary coordination sphere
because it is believed that such a group may play a role in proton
transfer in the process of proton reduction and hence improve
catalytic efficiency. Wherever the base groups are located in the
mimics, studies demonstrated that those groups were actively
involved in the protonation of the mimics.[4,14–18] An important
consideration is that the base group in a model complex may
initiate a proton-coupled electron transfer process in catalysis.
In natural systems, this is one of the key processes employed
*
Correspondence to: Xiaoming Liu, College of Biological, Chemical Sciences and
Engineering, Jiaxing University, Jiaxing 314001, China. E-mail: xiaoming.
liu@mail.zjxu.edu.cn
a Department of Chemistry, Nanchang University, Nanchang, 330031, China
b College of Biological, Chemical Sciences and Engineering, Jiaxing University,
Jiaxing, 314001, China
Appl. Organometal. Chem. 2013, 27, 253–260
Copyright © 2013 John Wiley & Sons, Ltd.

Y. Li et al.
of polyvinyl chloride (PVC), a plastic found in many applications in
industry and is available at low cost, to achieve artificial systems
to mimic the enzyme.[37] By doping the functionalized polymers
with multiwall carbon nanotubes, film electrodes were assem-
bled with the aid of Nafion.
Obviously, in addition to availability and low cost, an ideal
material into which a diiron mimic is to be incorporated should
exhibit flexibility in tailoring its properties, for example, solubility,
hydrophilicity, and protonability. Being inspired by our recent
success in exploiting the PVC plastic,[37] we turned to another
commercially available polymeric material – polyethyleneimine
(PEI) – which is an additive widely used in industry. Herein, we
report our exploration of the reactivity of the primary and
secondary amine groups (NH2 and NH, respectively) of PEI to
prepare PEI-based materials functionalized with diiron carbonyl
mimics. Through premodifying PEI with diphenylphosphinamine
(NPPh2), diiron models, [Fe2(m-S)2(CO)6] (1) or [Fe2(m-S2C2H4)(CO)
6] (3), were chemically grafted onto PEI skeleton via substitution
of the bound CO by the premodified NPPh2 groups. The resultant
materials were characterized using NMR, Fourier transform
infrared spectroscopy, and TGA techniques. Cyclic voltammetry
was employed to investigate the electrochemistry of the mate-
rials and their catalysis on proton reduction in DMF–acetic acid
medium. It was found that increasing the flexibility of the
catalytic center (the diiron unit) and immobilization of carboxylic
acid onto PEI could improve the electrocatalytic efficiency.
Figure 1. Schematic view of the H-cluster of [FeFe]-hydrogenase in its
rest state.
by nature to reduce the reaction energy barrier.[19–24] Transition
metal complexes bearing a pendant base functioning as a proton
relay are not unprecedented.[25,26] A recent report by Rauchfuss
and his coworkers demonstrated well that the azo group of the
bridging head of a diiron model helped activate hydrogen via
the PCET mechanism at a significantly lower potential compared
to an analogous system without such a base group.[20] This is
probably thus far the most concrete experimental evidence
providing rationalization for assigning the uncertain X ( Fig. 1) to
an NH group.
While the number of mimics of the diiron core, [Fe2(CO)6ꢀx
]
(x = 1, 2), has increased exponentially in recent years, studies to
address concerns beyond the diiron subunit, e.g. the peripheral
environment in which the H-cluster is harbored, have been
relatively rare. It is undoubted that the surrounding protein
domains of the active site of the enzyme are indispensable for
the enzyme to exhibit its catalytic function. This nonmetallic part
of the enzyme not only offers a safe guard for the fragile
and highly oxygen-sensitive active site but also functions as a
platform to construct a channel to shuttle the substrate in
and the product out. Nature devises the channel in such a
way that those processes are thermodynamically downhill to
avoid extra energy expenses in the catalysis. Jones and her
coworkers grafted a diiron carbonyl unit onto a short peptide
chain a few years ago, and this was probably the very first
attempt employing an approach analogous to nature to
address this issue.[27]
Experimental
Reagents and Equipment
Reactions and manipulations were performed using standard
Schlenk technique under argon atmosphere when necessary. All
solvents were dried following standard procedures prior to use
and stored under argon atmosphere. PEI (branched, Mw = 1800
and 600) and chlorodiphenylphosphine were purchased from Alfa
Aesar and triphenylphosphine from Sinopharm Chemical Reagent
Co., Ltd. General chemicals were purchased from local commercial
suppliers and further purified when necessary. Complexes [Fe2(m-S)
2(CO)6] (1), [Fe2(m-S)2(CO)5PPh3] (2), [Fe2(m-S2C2H4)(CO)6] (3), and
[Fe2(m-S2C2H4)(CO)5PPh3] (4) (refer to ESI) were synthesized follow-
ing published procedures.[41–43] Infrared (IR) spectra were recorded
on a Scimitar 2000 (Varian). NMR spectral data were collected on an
Avance 600 or 400 (Bruker) in CDCl3. TGA measurement was carried
out on a NETZSCH STA 409 PC/PG at a heating rate of 10ꢁC minꢀ1
under nitrogen purging. Electrochemical investigation was
performed under argon in a customized glass cell with a three-
electrode system as described elsewhere.[44] In the setup, the
working electrode was a glassy carbon disk (f = 5 mm), and a
glassy carbon strip electrode was used as a counter electrode.
The reference electrode was Ag/AgCl with an inner reference
Such a peripheral component is also useful in assembling an
electrode by using mimics. But developing mimicking systems of
this kind needs to draw further attention because only a handful
examples have been reported to date.[28–32] One of the difficulties
is the integration of the diiron mimics into such a peripheral
environment, which can be used as a scaffold in the assembly.
Selecting an appropriate scaffold is another nontrivial task. In the
reported systems, forming a monolayer of the mimics on the
surface of a vitreous carbon electrode was one of the approaches
to construct an electrode.[33,34] Further improvement was achieved
by anchoring diiron models into pro-polymerized polypyrrole film
via diffusion and then chemical reaction.[35,36] Recently, we
reported polymeric systems functionalized with diiron mimics
prepared either by polymerizing diiron models possessing alkynyl
group(s) [14,37,38] or by joining the model to an organic diazide via
‘click’ chemistry.[39] Those functionalized polymers were success-
fully employed to construct film electrodes that exhibited
electrochemical responses in the presence or absence of an acid.
However, the tedious, time-consuming, and perhaps costly nature
of their synthesis poses a disadvantage for their further develop-
ment. To bypass this hurdle, it is suggested that utilizing materials
commercially available may be an option in this regard. It was
reported that resin was successfully employed to support diiron
models.[40] Recently, we exploited the functionality of the C-Cl bond
solution of [NBut4]BF4 (0.45 mol lꢀ1) and [NBut4]Cl (0.05 mol lꢀ1
)
in dichloromethane (DCM). All potentials were quoted against
a ferrocenium/ferrocene couple.
Functionalization of PEI with Diphenylphosphine
PEI-P-A
To a solution of PEI (4.30 g) and anhydrous NEt3 (7 ml, 50 mmol)
in DCM (80 ml) cooled in an ice bath, ClPPh2 (9 ml, 50 mmol) in
DCM (20 ml) was slowly added dropwise under vigorous stirring.
wileyonlinelibrary.com/journal/aoc
Copyright © 2013 John Wiley & Sons, Ltd.
Appl. Organometal. Chem. 2013, 27, 253–260

[FeFe]-hydrogenase mimics
After the addition, the reaction was stirred for a further 4 h at
room temperature. Removal of the solvent under reduced
pressure gave an off-white solid, which was redissolved in DCM
(5 ml). To the solution was added hexane (500 ml) under stirring.
Upon the addition, precipitates formed instantly and suspended
in the solution. The mixture was left standing for 2 h to settle
the product at the bottom of the reaction vessel. A pale-yellow
solid (i.e. PEI-P-A) was collected by filtration, washed with
anhydrous ethanol, and then dried under vacuum to a
constant weight (10.6 g). IR (KBr): 3407 (NH), 3051, 1586, 1479,
1435 cmꢀ1 (benzene). 31P NMR (CDCl3, 600 MHz): d 22.01, 61.60,
62.21 ppm. Other functionalized polymers – PEIs, PEI-P-B,
PEI-P-B0, PEI-P-C, PEI-P-D, and PEI-P-E – were analogously
prepared (Table 1).
Preparation of the Polymer Functionalized with
Diiron Models
PEI-P-Fe2-A
PEI-P-A (2.00 g) in anhydrous THF (5 ml) was added to a solution
of complex 1 (2.54 g, 7.4 mmol) in anhydrous THF (70 ml) under
rapid stirring. After the addition, the reaction mixture was
refluxed and stirred for a further 24 h at 70ꢁC. When the reaction
was completed as monitored using thin-layer chromatography,
the mixture was filtered, and the filtrate was collected. Removal
of the solvent produced a residue, which was redissolved in
anhydrous DCM (5 ml). The solution was added to hexane
(500 ml) dropwise under vigorous stirring to precipitate the
product. The product (dark brown, 3.1 g) was collected via filtra-
tion, washed with anhydrous ethanol, and dried under reduced
pressure. The product was kept under Ar atmosphere for further
use. IR (KBr): 2054, 2013, 1995 1985, 1972, 1934 cmꢀ1 (CO). 31P
NMR (CDCl3, 600 MHz): d 32.97, 69.66, 71.45 ppm. PEI-P-Fe2-A:
found S 11.10, Fe 19.33. The other polymers functionalized with
diiron mimics were analogously prepared (Table 2).
PPh2-O-(CH2)11-Br
A solution of HO-(CH2)11-Br (2.24 g, 8.9 mmol) and Et3N (0.63 ml,
8.9 mmol) were mixed in anhydrous THF (50 ml) at 0ꢁC. To the
precooled solution, ClPPh2 (1.6ml, 8.9 mmol) in THF (2ml) was
slowly added dropwise under vigorous stirring. After the addition,
the reaction was stirred for a further 1 h at low temperature.
Removal of the solvent followed by purification using flash chroma-
tography (eluent, ethyl acetate/petroleum ether = 20%/80%)
produced a white solid (1.56 g, 40%). IR (KBr): 1592, 1485, 1439
Results and Discussion
Pre-functionalization of PEI with Diphenylphosphine and
Their Reaction with the Diiron Models
(benzene), 3058 (benzene C-H), 2929, 2855 (CH2) cmꢀ1 31P NMR
.
(CDCl3, 400 MHz): d 31.34 ppm.
It is well known that phosphine ligand substitutes readily the CO
of diiron hexacarbonyl complexes, and numerous mimics have
been prepared in this manner.[43,45–49] To functionalize PEI with
diiron mimics, we exploited this reactivity by first premodifying
PEI through its reaction with ClPPh2 and then reacting with diiron
hexacarbonyl complexes. In this study, four types of premodified
PEI were prepared (Table 1). PEI-P-A, PEI-P-B, PEI-P-C, PEI-P-D, and
PEI-P-E were prepared by reacting ClPPh2 with PEI of average
molecular weight 1800. These premodified PEIs differed in the
content of chemically anchored diphenylphosphine (Table 1).
The second type was PEI-P-B0, which was prepared using PEI with
a smaller molecular weight (600). To allow more flexibility for
immobilized diiron models, we linked the diphenylphosphine to
PEI via a long aliphatic chain. Reacting 11-bromoundecanol with
PEI-(CH2)11-O-P
PPh2-O-(CH2)11-Br (550 mg, 1.26 mmol), Et3N (1.3 mmol, 0.18 ml),
and PEI (143 mg) were reacted in DCM (10 ml). The reaction was
similarly performed and worked up as described earlier to
produce a yellow oily liquid (Table 1).
HOOCCH2-PEI-P
BrCH2COOH (278 mg, 2 mmol) was reacted with PEI-P-B (540 mg)
in DCM (20 ml) in the presence of anhydrous K2CO3 (300 mg,
2.18 mmol). The reaction was similarly performed and worked
up as described earlier to produce a white solid (Table 1).
Table 1. The ratio of (PEI) and chlorodiphenylphosphine (CDPP) in the preparation of functionalized PEIs, their physical properties, and 13P NMR
PEI-P
PEI (g)/ClPPh2 (mmol)
Physical properties
13P NMR (CDCl3)
Yield (g)
PEI-P-A
PEI-P-B
PEI-P-C
PEI-P-D
4.3/50
4.3/38
4.3/15
2.15/4
Pale-yellow solid
Pale-yellow solid
White solid
22.01, 61.60, 62.21
10.6
8.7
21.94, 62.16
21.40, 23.55, 35.39, 36.82, 40.78, 62.12
18.23, 21.89, 24.77, 29.95, 30.44, 36.09,
37.03, 40.94, 41.30, 62.68
4.1
Pale-yellow oily liquid
1.6
PEI-P-E
PEI-P-B0
PEI-(CH2)11-O-Pb
HOOCCH2-PEI-Pc
4.3/2
Pale-yellow oily liquid
18.46, 21.92, 24.12, 24.33, 24.58, 30.34, 36.03,
36.97, 41.06, 41.40, 62.82
0.50
a
4.3/38
4.3/38
Pale-yellow solid
Yellow viscous liquid
White solid
21.47, 35.38, 36.69, 40.06, 40.79, 61.90, 61.12, 61.67
31.05
2.52
0.24
0.23
17.06, 24.04, 27.60, 28.42, 29.86, 30.54, 32.32, 50.16
aPEI with a molecular weight of 600 was employed.
bNo 13P NMR data were available for PEI-(CH2)11-O-P, but 13P NMR of its precursor (PPh2-O-(CH2)11-Br) showed a signal at 1.34 ppm.
cPEI-P-B: BrCH2COOH = 0.54 g : 2 mmol and 1729 cmꢀ1 (KBr) for the carboxyl group.
Appl. Organometal. Chem. 2013, 27, 253–260
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/aoc

Y. Li et al.
ClPPh2 produced PPh2-O-(CH2)11-Br, which further reacted with
PEI to achieve the third premodified PEI, PEI-(CH2)11-O-P. In the
last premodified PEI (HOOCCH2-PEI-P), a carboxylic group was
introduced by reacting PEI-P-B with bromoacetic acid. An absorp-
tion band at 1729 cmꢀ1 of the resultant product indicated the
successful immobilization of the carboxyl group. All the prepara-
tions of PEI-Ps are illustrated in Schemes 1 and 2.
Polyethyleneimine is a hygroscopic liquid. As shown in Table 1,
the premodified PEIs were solid in most cases. When the content
of the chemically grafted NPPh2 was relatively low or a long
aliphatic chain was introduced, the pre-functionalized PEIs could
be oily liquids. These PEIs are soluble in common organic solvents
such as DMF, THF, and DCM. The premodification was confirmed
by IR (supporting information, Fig. S1) and 13P NMR spectroscopies
(supporting information, Fig. S2). The PEIs grafted with NPPh2
showed a number of signals of 13P NMR ranging from 15 to
65 ppm, which is in agreement with the multiple reaction sites of
PEI with ClPPh2 and the coupling between P and the N atoms of
the backbone.
Reaction of the PEIs premodified with NPPh2 (PEI-P) in THF
with the diiron complexes (1 or 3) led to the polymers function-
alized with diiron mimics (PEI-P-Fe2) in good yields in most cases
(Table 2). Unlike the PEI-Ps, all the PEI-P-Fe2s were dark solids and
soluble in organic solvents such as dichloromethane and DMF.
For the same reasons as for PEI-Ps (Table 1), they also showed
NH2
(CH2)2
H
N
N
N
N
m
n
H
(H2C)2
PEI
NH2
CH2Cl2
NEt3
ClPPh2
NH2
PPh2
N
(H2C)2
N
N
H
N
m
n
(H2C)2
NH
PEI-P
PPh2
Complex 1 or 3
THF
NH2
Ph2P
(H2C)2
N
N
N
H
N
m
n
(H2C)2
NH
PEI-P-Fe2
PHPh2
S
S
S
S
Fe
CO
CO
CO
=
or
Fe
OC
Fe
CO
Fe
OC
OC
CO
OC
CO
Scheme 1. Illustrative representations and preparation of the
premodified PEIs and their further functionalization with diiron models
(see Tables 1 and 2 for detailed denotations of the macromolecules).
wileyonlinelibrary.com/journal/aoc
Copyright © 2013 John Wiley & Sons, Ltd.
Appl. Organometal. Chem. 2013, 27, 253–260

[FeFe]-hydrogenase mimics
NH2
(CH2)2
N
H
N
N
H
N
m
n
(H2C)2
PEI
NH2
CH2Cl2
NEt3
Br(CH2)11PPh2
NH(CH2)11PPh2
(CH2)11PPh2
(H2C)2
N
N
N
N
m
m
m
n
(CH2)11PPh2
(H2C)2
NH
(CH2)11PPh2
PEI-(CH2)11-O-P
Figure 2. IR spectra of PEI-P-Fe2-A, -B, -C, -D, -E (KBr) and complex 2
(DCM).
NH2
PPh2
N
(H2C)2
N
N
H
N
n
(H2C)2
NH
PEI-P-B
PPh2
BrCH2COOH
CH2Cl2
H
N
PPh2
N
(H2C)2
N
CH2COOH
N
N
n
CH2COOH
(H2C)2
NH
PPh2
HOOCCH2-PEI-P
Figure 3. IR spectra of PEI-P-Fe2-B, PEI-P-Fe2-B0, PEI-P-Fe2-B-2, PEI-P-Fe2-B0-
2, PEI-(CH2)11-O-P-Fe2-B, HOOCCH2-PEI-P-Fe2 (KBr).
Scheme 2. Illustrative representations and preparation of PEI-(CH2)11-O-P
and HOOCCH2-PEI-P (see Table
1 for detailed denotations of the
premodified PEIs).
suggested that the mimics grafted onto the PEI skeleton via the
diphenylphosphine were maintained as a diiron unit.
multi-signals in their 13P NMR spectra (Table 2 and supporting
information, Fig. S3). Recently, we reported a number of polymers
functionalized with diiron mimics.[14,37–39] Those polymers grafted
with either diiron hexacarbonyl units or diiron pentacarbonyl units
all exhibited characteristic spectral profiles corresponding to the
grafted diiron mimics, although broadening was observed because
of a polymeric effect. However, for the polymers reported in this
work, the IR spectra are more complicated (Figs 2 and 3). By
comparing the spectra of the control complexes (2 and 4, Table 2),
it is obvious that the diiron pentacarbonyl units are dominant.
However, a sharp absorption band at about 2020 cmꢀ1 suggests
that further substitution may be possible to form diiron
tetracarbonyl units.[43] Either a free NPPh2 group or N atom from
the PEI backbone could complete the second substitution.
Therefore, the multiplicity of IR spectra of the polymers is caused
not only by the polymeric effect but also by the existence of
diiron tetracarbonyl units.
Thermal Stabilities of PEI-Ps and PEI-P-Fe2s
The thermal stability of the polymer was evaluated using TGA
between 298 and 1073 K under a steady flow of nitrogen. The
thermal decomposition of the polymers is shown in Fig. 4. For
comparison, TGA of PEI-P was also performed. Like in the
case of the polymers with diiron mimics that we reported
before,[14,37–39,43] attaching the diiron mimics onto the polymers
compromises thermal stability. By comparison with the thermal
gravimetric trace of PEI-P, the first stage of thermal decomposition
of the functionalized polymers was attributed to the diiron
carbonyl units. This assignment was further supported by examin-
ing the IR absorption bands of CO after the polymers were heated
to about 573 K. The residue obtained from heating did not show
any CO absorption bands. Compared with analogous functional-
ized polymers that we reported previously,[38,39] the starting
temperature of decomposition was evidently higher. This showed
that the functionalized polymers with PEI backbone have better
thermal stability. The improvement in thermal stability could be
attributed to intramolecular and intermolecular hydrogen interac-
tions. Furthermore, flexible PEI chains could also offer a shielding
effect against thermal decomposition.[14]
Analyzing iron content suggests a maximum iron content of
~20% and lowest iron content of ~5% (Table 2). It seems that
increasing the content of ClPPh2 in the preparation of PEI-Ps
did not proportionally increase the diiron mimics on the PEI-P
backbone, but the molar ratios between the Fe and S were
calculated as 0.997 ꢂ 0.006: almost a perfect value of 1:1. This
Appl. Organometal. Chem. 2013, 27, 253–260
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/aoc

Y. Li et al.
Figure 4. (a) TGA traces of the polymers PEI-P-Fe2-A, -B, -C, -D, -E and
PEI-P recorded in nitrogen flow (heating rate 10 ꢁC minꢀ1). (b) TGA
traces of the polymers PEI-P-Fe2-B, PEI-P-Fe2-B0, PEI-P-Fe2-B-2, PEI-P-Fe2-B0-
2, PEI-(CH2)11-O-P-Fe2-B, HOOCCH2-PEI-P-Fe2 and PEI-P recorded in nitrogen
flow (heating rate 10ꢁC minꢀ1).
Figure 5. (a) Cyclic voltammograms of PEI-P-Fe2-A (28.6 mg in 3.5 mL)
PEI-P-Fe2-B (7.9 mg in 3.5 mL), PEI-P-Fe2-C (9.0 mg in 3.5 mL), PEI-P-Fe2-D
(10 mg in 3.5 mL), PEI-P-Fe2-E (29 mg in 3.5 mL), PEI-P(7.9 mg in 3.5 mL)
and complex 2 (C = 4.79 ꢃ 10ꢀ3 mol Lꢀ1) in 0.1 mol Lꢀ1 [NBu4]BF4/DMF
at a scanning rate of 0.1 V sꢀ1 under Ar atmosphere at 298 K. (b) Cyclic
voltammograms of PEI-P-Fe2-B (7.9 mg, 3.5 mL), PEI-P-Fe2-B0 (9 mg,
3.5 mL), PEI-P-Fe2-B-2 (8.9 mg, 3.5 mL), PEI-P-Fe2-B0-2 (7.5 mg, 3.5 mL),
PEI-(CH2)11-O-P-Fe2-B (8.5 mg, 3.5 mL), HOOCCH2-PEI-P-Fe2 (7.9 mg,
3.5 mL) in mol Lꢀ1 [NBu4]BF4/DMF at a scan rate of 0.1 V sꢀ1 under Ar
atmosphere at 298 K.
Electrochemistry of the Functionalized Polymers, PEI-P-Fe2s
and Their Catalysis on Proton Reduction
Because the functionalized polymers were not sufficiently soluble
in acetonitrile, electrochemistry of the functionalized polymers
was performed in DMF. Because our main concern was the reduc-
tion process and its catalysis on proton reduction, oxidation of
these polymers was not examined. Cyclic voltammograms (CVs)
of the functionalized polymers and reduction potentials are
shown in Fig. 5 and Table 2, respectively. In the CVs shown in
Fig. 5, the normalized current (J, mA mgꢀ1) is defined as current
(i, mA) divided by iron content (mg) calculated using the iron
percentage (Fe%, Table 2) to eliminate the influence of the varia-
tion in content of the chemically grafted diiron mimics. As shown
in the CVs, the electrochemical behavior of these polymers was
affected by the content of diiron mimics, the presence of carboxylic
acid (HOOCCH2-PEI-P-Fe2), the distance of the diiron mimics to the
PEI backbone (PEI-(CH2)11-O-P-Fe2), and the diiron mimics.
The variation in average molecular weight of PEI did not seem
to exert any influence on the electrochemistry (PEI-P-Fe2-B and
PEI-P-Fe2-B0). Conversely, the diiron hexacarbonyl precursor
showed a significant influence on the resultant polymers. Reac-
tion of complex 3 with PEI-P resulted in polymers PEI-P-Fe2-B-2
and PEI-P-Fe2-B0-2, which possessed diiron units analogous to
complex 4. Electrochemically, they behaved very much like
complex 4 (Table 2). When complex 1 was used, however, the
electrochemical responses of the resultant polymers (PEI-P-Fe2-A
to PEI-P-Fe2-E) were much more complicated (Fig. 6a, b). The most
distinctive feature of their electrochemistry, as shown in Fig. 5, was
that a few reduction processes appeared at more positive poten-
tials, in addition to the process analogous to that of complex 2 at
about ꢀ1.4V. These additional processes were not observed when
the content of diiron mimics decreased (PEI-P-Fe2-E). As indicated
by its reduction potential and IR spectral absorption bands (Table 2),
complex 1 and hence 2 are stronger nucleophiles than complexes
3 and 4, respectively. Considering the electron-withdrawing
nature of the diiron units in the polymers, supposedly mostly
pentacarbonyl, the reduction potential of a diiron unit would
be driven positively by its surrounding diiron units. Such an
effect would certainly be less important when the diiron content
decreased (PEI-P-Fe2-E), and therefore, its reduction behaved
more simply compared with those of the others (Fig. 5a). Increas-
ing the distance of the diiron units to the PEI backbone did not
show any drastic change in electrochemistry, as shown by the
CV of polymer PEI-(CH2)11-O-P-Fe2-B (Fig. 5b).
To examine the catalytic behavior of the functionalized
polymers, we performed their electrochemical catalysis on pro-
ton reduction in DMF in the presence of acetic acid. As already
wileyonlinelibrary.com/journal/aoc
Copyright © 2013 John Wiley & Sons, Ltd.
Appl. Organometal. Chem. 2013, 27, 253–260

[FeFe]-hydrogenase mimics
content show better catalytic performance. There is no straight-
forward explanation for this. One possible cause is that evolution
of one hydrogen molecule via catalytic reduction of proton
involves the transfer of two proton and electrons stepwise,
respectively. Synergistic actions among the diiron units could
enhance the catalysis and, therefore, high-density polymers
(high iron content) showed better catalytic efficiency.
Interesting results are further found in Fig. 6(b). As indicated by
the normalized current at 70 ml of the acid, either varying PEI
(average molecular weight of 1800 and 600, respectively) or
diiron complex (1 and 3, respectively) in the functionalization of
the polymers did not significantly improve the catalytic perfor-
mance. The other two functionalized polymers – PEI-(CH2)11-O-P-
Fe2-B and HOOCCH2-PEI-P-Fe2 – however, showed improvement
in catalytic performance. In particular, the latter was derived using
the same PEI-P-B precursor as for the preparation of PEI-P-Fe2-B,
but its normalized current increased by over one-fold. For
PEI-(CH2)11-O-P-Fe2-B, the long aliphatic chain, -(CH2)11-,
increases the flexibility of the catalytic site (diiron center). This
flexibility could be beneficial to the catalysis because the
involvement of transfer of multiple protons and electrons per
hydrogen molecule requires structural rearrangement of the
catalyst during catalysis. The results show that the basic groups
of PEI help in improving catalytic efficiency by taking part in
proton transfer. Carboxylic acid ought to be a better proton donor
than the protonated amine group. Therefore, it is not surprising
that the functionalized polymer, HOOCCH2-PEI-P-Fe2, exhibited
much higher catalytic efficiency than PEI-P-Fe2-B.
Conclusions
Figure 6. (a) Curve of PEI-P-Fe2-A (5.8mg in 3.5mL), PEI-P-Fe2-B (7.9mg
in 3.5mL), PEI-P-Fe2-C (9mg in 3.5mL), PEI-P-Fe2-D (10 mg in 3.5mL),
In this study, we employed PEI to construct a mimicking system
for [FeFe]-hydrogenase. PEI was not only used as a scaffold to
support diiron mimics but also its basicity was exploited to
improve proton transfer in catalysis. Electrochemical investiga-
tion into the seven functionalized polymers suggests that the
amine groups of PEI played a role in increasing catalytic efficiency
by improving proton transfer. Further enhancement in proton
transfer was achieved by introducing carboxylic acid onto the
skeleton of PEI, and the catalytic performance was improved by
over one-fold. By dangling the catalytic unit onto the backbone
of PEI via a long aliphatic chain, catalytic efficiency was also
improved because of the increase of flexibility of the catalytic
units (diiron mimics). The flexibility allows a feasible structural
rearrangement of the catalytic center, which during catalysis, is
required by the catalytic process involving multiple proton and
electron transfers for the evolution of each hydrogen molecule.
Compared with the system based on PVC we reported recently,[37]
the PEI-derived functionalized polymers enjoy several advantages:
(i) These polymers are highly hydrophilic. (ii) The rich primary and
secondary amine groups allow diverse functionalization to further
tailor the PEI-based systems, for example, immobilizing quantum
dot to construct a photocatalytic system or functionalizing with
an electron mediator to enhance electron transfer. Thus, utilizing
PEI to develop artificial systems inspired by [FeFe]-hydrogenase
deserves further exploration.
PEI-P-Fe2-E (29 mg in 3.5 mL) and complex 2 (C= 4.79 ꢃ 10ꢀ3 mmolmLꢀ1
)
in 0.1molLꢀ1 [NBu4]BF4/DMF at a scanning rate of 0.1Vsꢀ1 under Ar atmo-
sphere at 298 K. (b) Curve of PEI-P-Fe2-B (7.9mg, 3.5mL), PEI-P-Fe2-B0
(8.6 mg, 3.5 mL), PEI-P-Fe2-B-2 (8.9 mg, 3.5 mL), PEI-P-Fe2-B0-2 (7.5 mg,
3.5 mL), PEI-(CH2)11-O-P-Fe2-B (8.5 mg, 3.5 mL), HOOCCH2-PEI-P-Fe2
(5.7 mg, 3.5mL) and complex 2 (4.79 ꢃ 10ꢀ3 mmolmLꢀ1) in 0.1mol Lꢀ1
[NBu4]BF4/DMF at a scanning rate of 0.1V sꢀ1 under Ar atmosphere at 298K.
mentioned in the discussion of their electrochemistry, catalytic
currents were normalized by dividing the current by the iron
content. All CVs of the catalysis are shown in the supporting
information, Figs S4 and S5. As shown by the CVs, all the func-
tionalized polymers showed no substantial catalysis in response
to the reduction processes of potentials more positive than
about ꢀ1.8 V in the presence of acetic acid. Instead, a catalytic
process was often observed at about ꢀ2.2 V. In some cases,
catalytic current increased so strongly that no wave-shaped
process could be observed. To compare the catalytic efficiency
of the functionalized polymers, we plotted normalized current
(J, mA mgꢀ1) against added acetic acid. Considering that not in
all cases could a peak catalytic current be read, a current at
ꢀ2.17 V in all cases was read and divided by the iron content
in the sample. For reasons of clarity and categorization, plots of
normalized current J versus acetic acid volume (ml) for polymers
PEI-P-Fe2-A to PEI-P-Fe2-E and the others are shown in Fig. 6(a)
and (b), respectively.
Supporting information
The normalized current versus the volume of the acid added
as shown in Fig. 6(a) suggests that the functionalized polymers
(PEI-P-Fe2-A, PEI-P-Fe2-C, and PEI-P-Fe2-D, Table 2) with high iron
Supporting information may be found in the online version of
this article.
Appl. Organometal. Chem. 2013, 27, 253–260
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/aoc

Y. Li et al.
[22] F. Quentel, G. Passard, F. Gloaguen, Energy Environ. Sci. 2012, 5, 7757.
[23] P. E. M. Siegbahn, M. R. A. Blomberg, Chem. Rev. 2010, 110, 7040.
[24] Y. A. Small, D. L. DuBois, E. Fujita, J. T. Muckerman, Energy Environ. Sci.
2011, 4, 3008.
[25] D. L. DuBois, R. M. Bullock, Eur. J. Inorg. Chem. 2011, 1017.
[26] M. R. DuBois, D. L. DuBois, Chem. Soc. Rev. 2009, 38, 62.
[27] S. Roy, S. Shinde, G. A. Hamilton, H. E. Hartnett, A. K. Jones, Eur. J.
Inorg. Chem. 2011, 1050.
Acknowledgments
We thank the Natural Science Foundation of China (Grant Nos.
20871064, 21171073), the Ministry of Science and Technology
(China) (973 program, Grant No. 2009CB220009), and the Gov-
ernment of Zhejiang Province (Qianjiang professorship for XL)
for supporting this work.
[28] D. Svedruzic, J. L. Blackburn, R. C. Tenent, J.-D. R. Rocha, T. B. Vinzant,
M. J. Heben, P. W. King, J. Am. Chem. Soc. 2011, 133, 4299.
[29] F. Zipoli, R. Car, M. H. Cohen, A. Selloni, J. Am. Chem. Soc. 2010, 132,
8593.
References
[1] D. J. Evans, C. J. Pickett, Chem. Soc. Rev. 2003, 32, 268.
[2] J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt, Science 1998,
282, 1853.
[3] Y. Nicolet, C. Piras, P. Legrand, C. E. Hatchikian, J. C. Fontecilla-Camps,
Struct. Fold. Des. 1999, 7, 13.
[30] E. Lojou, X. Luo, M. Brugna, N. Candoni, S. Dementin, M. T.
Giudici-Orticoni, J. Biol. Inorg. Chem. 2008, 13, 1157.
[31] T. Yamaguchi, S. Masaoka, K. Sakai, Chem. Lett. 2009, 38, 434.
[32] K. A. Vincent, A. Parkin, F. A. Armstrong, Chem. Rev. 2007, 107, 4366.
[33] C. J. Pickett, S. K. Ibrahim, D. L. Hughes, Faraday Discuss. 2000, 116,
235.
[4] A. S. Pandey, T. V. Harris, L. J. Giles, J. W. Peters, R. K. Szilagyi, J. Am.
Chem. Soc. 2008, 130, 4533.
[34] V. Vijaikanth, J. F. Capon, F. Gloaguen, P. Schollhammer, J. Talarmin,
Electrochem. Commun. 2005, 7, 427.
[35] S. Ibrahim, P. M. Woi, Y. Alias, C. J. Pickett, Chem. Commun. 2010, 46,
8189.
[5] P. Volkers, T. B. Rauchfuss, J. Inorg. Biochem. 2007, 101, 1748.
[6] C. A. Boyke, T. B. Rauchfuss, S. R. Wilson, M. M. Rohmer, M. Benard,
J. Am. Chem. Soc. 2004, 126, 15151.
[7] F. Gloaguen, J. D. Lawrence, M. Schmidt, S. R. Wilson, T. B. Rauchfuss,
J. Am. Chem. Soc. 2001, 123, 12518.
[36] S. K. Ibrahim, X. M. Liu, C. Tard, C. J. Pickett, Chem. Commun. 2007,
1535.
[8] J. D. Lawrence, T. B. Rauchfuss, S. R. Wilson, Inorg. Chem. 2002,
41, 6193.
[9] Y. C. Liu, L. K. Tu, T. H. Yen, G. H. Lee, S. T. Yang, M. H. Chiang, Inorg.
Chem. 2010, 49, 6409.
[10] M. T. Olsen, M. Bruschi, L. De Gioia, T. B. Rauchfuss, S. R. Wilson, J. Am.
Chem. Soc. 2008, 130, 12021.
[37] L. Wang, Z. Xiao, X. Ru, X. Liu, Rsc Adv. 2011, 1, 1211.
[38] X. Ru, X. Zeng, Z. Li, D. J. Evans, C. Zhan, Y. Tang, L. Wang, X. Liu,
J. Polymer Sci., Part A: Polymer Chem. 2010, 48, 2410.
[39] C. Zhan, X. Wang, Z. Wei, D. J. Evans, X. Ru, X. Zeng, X. Liu, Dalton
Trans. 2010, 39, 11255.
[40] K. N. Green, J. L. Hess, C. M. Thomas, M. Y. Darensbourg, Dalton Trans.
[11] S. Tschierlei, S. Ott, R. Lomoth, Energy Environ. Sci. 2011, 4, 2340.
[12] C. Tard, C. J. Pickett, Chem. Rev. 2009, 109, 2245.
[13] X. M. Liu, S. K. Ibrahim, C. Tard, C. J. Pickett, Coord. Chem. Rev. 2005,
249, 1641.
2009, 4344.
[41] J. L. Stanley, T. B. Rauchfuss, S. R. Wilson, Organometallics 2007, 26, 1907.
[42] E. J. Lyon, I. P. Georgakaki, J. H. Reibenspies, M. Y. Darensbourg,
J. Am. Chem. Soc. 2001, 123, 3268.
[14] X. Liu, X. Ru, Y. Li, K. Zhang, D. Chen, Int. J. Hydrogen Energ. 2011, 36, 9612.
[15] W. Zhong, Y. Tang, G. Zampella, X. F. Wang, X. L. Yang, B. Hu, J. Wang,
Z. Y. Xiao, Z. H. Wei, H. W. Chen, L. D. Gioia, X. M. Liu, Inorg. Chem.
Commun. 2010, 13, 1089.
[43] Y. Wang, Z. Li, X. Zeng, X. Wang, C. Zhan, Y. Liu, X. Zeng, Q. Luo,
X. Liu, New J. Chem. 2009, 33, 1780.
[44] X. H. Zeng, Z. M. Li, Z. Y. Xiao, Y. W. Wang, X. M. Liu, Electrochem.
Commun. 2010, 12, 342.
[16] Z. Xiao, Z. Li, G. Qian, L. Long, Asian J. Chem. 2012, 24, 249.
[17] M. H. Cheah, C. Tard, S. J. Borg, X. Liu, S. K. Ibrahim, C. J. Pickett,
S. P. Best, J. Am. Chem. Soc. 2007, 129, 11085.
[18] S. Lounissi, J. F. Capon, F. Gloaguen, F. Matoussi, F. Y. Petillon,
P. Schollhammer, J. Talarmin, Organometallics 2010, 29, 1296.
[19] J. M. Camara, T. B. Rauchfuss, J. Am. Chem. Soc. 2011, 133, 8098.
[20] M. T. Olsen, T. B. Rauchfuss, S. R. Wilson, J. Am. Chem. Soc. 2010, 132, 17733.
[21] P. Poddutoori, D. T. Co, A. P. S. Samuel, C. H. Kim, M. T. Vagnini,
M. R. Wasielewski, Energy Environ. Sci. 2011, 4, 2441.
[45] J. Hou, X. J. Peng, Z. Y. Zhou, S. G. Sun, X. Zhao, S. Gao, J. Organomet.
Chem. 2006, 691, 4633.
[46] A. K. Justice, G. Zampella, L. De Gioia, T. B. Rauchfuss, J. I. van der
Vlugt, S. R. Wilson, Inorg. Chem. 2007, 46, 1655.
[47] C. M. Thomas, O. Rudiger, T. Liu, C. E. Carson, M. B. Hall,
M. Y. Darensbourg, Organometallics 2007, 26, 3976.
[48] N. Wen, F. F. Xu, Y. A. Feng, S. W. Du, J. Inorg. Biochem. 2011, 105,
1123.
[49] X. H. Zeng, Z. M. Li, X. M. Liu, Electrochim. Acta 2010, 55, 2179.
wileyonlinelibrary.com/journal/aoc
Copyright © 2013 John Wiley & Sons, Ltd.
Appl. Organometal. Chem. 2013, 27, 253–260

Featured Compounds
1
3
2
Compound Details
Structure Search
Compound Details
Structure Search
Compound Details
Structure Search
FC-1