Design of an Os Complex-Modified Hydrogel with Optimized Redox Potential for Biosensors and Biofuel Cells

Article abstract of DOI:10.1002/chem.201504591

Multistep synthesis and electrochemical characterization of an Os complex-modified redox hydrogel exhibiting a redox potential ≈+30 mV (vs. Ag/AgCl 3 m KCl) is demonstrated. The careful selection of bipyridine-based ligands bearing N,N-dimethylamino moieties and an amino-linker for the covalent attachment to the polymer backbone ensures the formation of a stable redox polymer with an envisaged redox potential close to 0 V. Most importantly, the formation of an octahedral N6-coordination sphere around the Os central atoms provides improved stability concomitantly with the low formal potential, a low reorganization energy during the Os^3+/2+ redox conversion and a negligible impact on oxygen reduction. By wiring a variety of enzymes such as pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenase, flavin adenine dinucleotide (FAD)-dependent glucose dehydrogenase and the FAD-dependent dehydrogenase domain of cellobiose dehydrogenase, low-potential glucose biosensors could be obtained with negligible co-oxidation of common interfering compounds such as uric acid or ascorbic acid. In combination with a bilirubin oxidase-based biocathode, enzymatic biofuel cells with open-circuit voltages of up to 0.54 V were obtained.

Full text of DOI:10.1002/chem.201504591

DOI: 10.1002/chem.201504591
Full Paper
&
Biosensors
Design of an Os Complex-Modified Hydrogel with Optimized
Redox Potential for Biosensors and Biofuel Cells
Piyanut Pinyou+,[a] Adrian Ruff+,[a] Sascha Pçller,[a] Su Ma,[b] Roland Ludwig,[b] and
Wolfgang Schuhmann*[a]
Abstract: Multistep synthesis and electrochemical characteri-
zation of an Os complex-modified redox hydrogel exhibiting
a redox potential ꢀ +30 mV (vs. Ag/AgCl 3m KCl) is demon-
strated. The careful selection of bipyridine-based ligands
bearing N,N-dimethylamino moieties and an amino-linker for
the covalent attachment to the polymer backbone ensures
the formation of a stable redox polymer with an envisaged
redox potential close to 0 V. Most importantly, the formation
of an octahedral N6-coordination sphere around the Os cen-
tral atoms provides improved stability concomitantly with
the low formal potential, a low reorganization energy during
the Os3+/2+ redox conversion and a negligible impact on
oxygen reduction. By wiring a variety of enzymes such as
pyrroloquinoline quinone (PQQ)-dependent glucose dehy-
drogenase, flavin adenine dinucleotide (FAD)-dependent glu-
cose dehydrogenase and the FAD-dependent dehydrogen-
ase domain of cellobiose dehydrogenase, low-potential glu-
cose biosensors could be obtained with negligible co-oxida-
tion of common interfering compounds such as uric acid or
ascorbic acid. In combination with a bilirubin oxidase-based
biocathode, enzymatic biofuel cells with open-circuit voltag-
es of up to 0.54 V were obtained.
Introduction
of these complexes can be tuned by the introduction of appro-
priate electron-withdrawing or electron-donating groups into
the ligand sphere.[15,18,20–22]
Wiring of enzymes to electrode surfaces by means of redox hy-
drogels[1] is of particular interest for the design of biosen-
sors,[2–4] enzymatic biofuel cells[5–7] or biophotovoltaic devi-
ces.[8–10] The electron transport between the electrode and the
prosthetic group of the enzyme within these bioactive films is
mediated by redox relays that are attached to the hydrogel
backbone. Osmium complexes of the type [Os(bpy)2Lx] (bpy=
2,2’-bipyridine, L=chloro ligand or ligand bearing a linker
group for the attachment to the polymer backbone; x=1, 2)
are widely used as redox mediators in hydrogels because of
the fast electron transfer rates they exhibit between the Os
centers and the prosthetic groups in various enzymes, for ex-
ample, among others glucose oxidase,[11–13] PQQ-dependent
glucose dehydrogenase (PQQ-sGDH, PQQ=pyrroloquinoline
quinone),[14,15] cellobiose dehydrogenase (CDH),[16,17] pyranose
dehydrogenase[18] or laccase.[19] Moreover, the formal potential
For glucose sensors, low operating potentials, which are de-
termined by the redox potential of the used mediator, are pre-
ferred to avoid interferences caused by electrochemically in-
duced side reactions, for example, the oxidation of ascorbate
or urate.[23,24] Moreover, low potential anodes ensure a high
open-circuit voltage in glucose-based biofuel cells, which is de-
termined by the difference between the potentials of the bioa-
node and the biocathode.[25]
One strategy to synthesize Os-based redox mediators with
low formal potentials is the introduction of chloro ligands to
the inner ligand sphere of the Os-complexes (L=Cl). Following
this approach, redox potentials <0 V (vs. Ag/AgCl 3m KCl) can
be achieved. However, the chloro ligands are labile and tend
to undergo ligand exchange reactions with solvent molecules,
the electrolyte, or functional groups (e.g., carbonyl or carboxyl
groups) at the polymer backbone. Moreover, unintended
ligand exchange reactions can be induced electrochemically
upon oxidation or reduction of the transition-metal center.
This effect was used for electrochemical induced crosslinking
of the polymer chains that bear free pyridine ligands to pro-
duce stable and dense films at the electrode surface.[26,27] How-
ever, these active layers contain Os complexes with different
ligand spheres and hence different formal potentials if the
ligand exchange is not complete and/or Os complexes with
different ligands are formed.[26] Moreover, the electron diffusion
coefficient within such polymers is low as compared to redox
hydrogels that bear redox relays bound through flexible link-
ers.[28,29]
[a] P. Pinyou,+ Dr. A. Ruff,+ Dr. S. Pçller, Prof. Dr. W. Schuhmann
Analytical Chemistry
Center for Electrochemical Sciences (CES)
Ruhr-Universität-Bochum, Universitätsstrasse 150
44780 Bochum (Germany)
E-mail: wolfgang.schuhmann@rub.de
[b] Dr. S. Ma, Dr. R. Ludwig
Department of Food Sciences and Technology
Vienna Institute of Biotechnology, BOKU - University of Natural Resources
and Life Sciences, Muthgasse 11/1/56; 1190, Vienna (Austria)
[+] These authors contributed equally to this work.
Supporting information and the ORCID identification number(s) for the au-
thor(s) of this article can be found under http://dx.doi.org/10.1002/
chem.201504591.
Chem. Eur. J. 2016, 22, 5319 – 5326
5319
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Other approaches for the formation of low potential Os
complexes use the introduction of NO ligands, that is, imida-
zole derivatives that contain ester groups. The coordination
takes place through the O atom of the carbonyl group and the
imino N atom in the imidazole ring.[15,17,30] However, the coordi-
native OÀOs bond is labile and again ligand exchange reac-
tions with solvent molecules[31] and functional groups within
the polymer chain may occur. This leads again to species with
a variety of different redox potentials.[15,17] Thus, an Os complex
with six N coordination sites would be preferable.
Results and Discussion
Synthesis of the Os complex-based redox mediator
The use of amino-modified bpy ligands ensures low redox po-
tentials.[20,29] However, the free amino groups in the outer
sphere of such ligands may react randomly with epoxy func-
tions at the polymer backbone. This leads finally to a number
of various products with different degree of binding (i.e.,
number of reacted amino groups). In addition, the large
number of ÀNH2 groups at the Os complex may induce unin-
tended crosslinking reactions between different polymer
chains already formed during the polymer modification pro-
cess. Furthermore, the ÀNH2 groups attached to bpy tend to
undergo oxidation even at ambient conditions, which would
cause to a change in the redox potential of the polymer-
bound Os complex. Dimethyl amino groups (ÀNMe2) provide
an even stronger donor effect than the ÀNH2 group (+I effect
of the methyl groups). Hence, bpy ligands carrying ÀNMe2
groups instead of ÀNH2 groups should ensure a low redox po-
tential of the corresponding Os complex. Moreover, the ÀNMe2
functions are inactive during the attachment of the complex
to the polymer backbone through a ring-opening reaction of
the polymer-bound epoxy groups at mild conditions. Anchor-
ing of the Os complex to the polymer chain is only due to the
reaction between the predefined binding site at the polymer
backbone and a corresponding linker species (see below). Fur-
thermore, N,N’-dimethylaniline species are inert against oxida-
tion through molecular oxygen under ambient conditions.[35]
A suitable ÀNMe2-modified bpy was synthesized starting
from N,N-dimethylamino pyridine (DMAP, 1, Scheme 1). The
pyridine precursor DMAP was first treated with [Sn(nBu)3Cl] to
form the pyridyl stannate 2. Treatment of 1 with nBuLi, fol-
lowed by treatment with I2 leads to halide 3. Finally, the Pd0-
catalyzed Stille cross-coupling between stannate 2 and halide
3 gives the ÀNMe2-modified bpy ligand dmabpy in 76%
yield.[36]
Osmium complexes with ligand spheres based on N,N’-dia-
lkylated-2,2’-diimidazole derivatives exhibit also remarkable
low redox potentials (<À150 mV vs. Ag/AgCl 3m KCl).[15,29]
However, these complexes reveal redox potentials that invoke
the catalytic reduction of oxygen by the polymer-bound redox
mediator itself (< +70 mV vs. Ag/AgCl 3m NaCl).[32] Hence, the
application in biosensors and biofuel cells that are operated
under ambient conditions avoiding parasitic reduction currents
at the bioanode is hampered.
Recently, we described the synthesis of an Os complex bear-
ing two bipyridyl (bpy) ligands and an ethylenediamine-modi-
fied bpy ligand for the modification of a poly(methacrylate)-
based polymer backbone by reacting epoxy groups introduced
in the polymer backbone and the ÀNH2 group at the bpy de-
rivative.[33] This redox polymer revealed a high chemical stabili-
ty even under harsh conditions (pH 3). However, the polymer
showed a rather high redox potential of +0.64 V versus Ag/
AgCl 3m KCl.[33]
To decrease the potential of the Os3+/2+ couple, electron-do-
nating groups need to be introduced at the bpy-based ligand
system.[22] The synthesis of such low-potential Os complexes
were first reported by Heller et al. in 2003[20,29] and later opti-
mized by Leech et al.[21,34] The authors introduced amino
groups in the para positions to the coordinating N-atoms in
the bpy ligands. Os complexes bearing two diamino-modified
bpy ligands and two chloro ligands show a formal potential of
À0.57 V (vs. Ag/AgCl 3m KCl).[21] Obviously, the introduction of
electron-donating groups leads to a strong decrease in the
redox potential of the corresponding Os complexes. However,
these complexes were decomposed by oxidation of the intro-
duced amino moieties.
The Os2+-dichloro complex
5 is obtained by treating
dmabpy and K2OsCl6 in dimethylformamide (DMF) at high
temperatures followed by the reduction of the Os species to
Os2+ with Na2S2O4 in water (Scheme 1).[37] Cyclic voltammo-
In this paper, we suggest the design of a new, highly-stable,
and low-potential Os complex overcoming limitations such as
unstable ligand sphere, unstable electron-donating substitu-
ents, as well as catalytic activity for oxygen reduction or inter-
ference oxidation. A N6-coordination sphere is envisaged,
avoiding any labile chloro or oxygen coordination. Dimethyl-
amino groups are used as stable electron-donating groups in
para position to the N-coordination sites of bpy ligands. One
of the bpy ligands additionally carries an amino-terminated
flexible linker that allows binding to a copolymer backbone for
the formation of the corresponding redox hydrogels. By inte-
gration of a variety of enzymes such as PQQ-dependent glu-
cose dehydrogenase, FAD-dependent glucose dehydrogenase,
and the FAD-containing dehydrogenase domain of cellobiose
dehydrogenase, low-potential glucose biosensors and bio-
anodes for biofuel cells were obtained and characterized.
Scheme 1. Synthesis of Os complex 5 [Os(dmabpy)2Cl2] by treating dmabpy
with K2OsCl6 in ethylene glycol. The dmabpy-ligand was synthesized
through the cross-coupling of stannate 2 and the iodo compound 3 starting
from DMAP (1) based on protocols described in ref. [36].
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5320
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. Voltammetric characterization of Os complexes 5 (a) and 12 (b) as
well as the redox polymer P-12 (c and d); a) cyclic voltammograms in the
potential window of À1.1 to À0.2 V (vs. Ag/AgCl 3m KCl) of 5 (c) in 0.1m
KCl/water drop-coated from EtOH onto a GC electrode (diameter=3 mm);
the dashed line corresponds to the I/E curve recorded with the bare GC elec-
trode; scan rate=100 mVsÀ1; b) cyclic voltammogram of 12 in 0.15m
NBu4PF6/MeCN at a Pt electrode (1 mm), scan rate=50 mVsÀ1c: cyclic vol-
tammogram of P-12 drop-cast onto a graphite electrode (3.05 mm) in PBS
buffer (pH 7.4); d) differential pulse voltammogram of P-12 in 3m KCl/water
deposited onto a Pt electrode (1 mm); pulse amplitude=15 mV, pulse
width=0.06 s
Scheme 2. Multistep synthesis of complex 12 starting from DMAP and acetal
6 (see the Supporting Information for synthesis of 6). The target compound
is obtained as the hexafluorophosphate salt through a metathesis reaction
with NH4PF6 in water.
range. However, as expected, due to the absence of any elec-
tron-donating groups at the bpy ligands, the redox potential
of the [Os(bpy)2(bpyamino)]2+ complex is +0.64 V versus Ag/
AgCl 3m KCl (note, the redox potential of [Os(bpy)2Cl2] is
ꢀ0 V).
grams in the potential range of À1.1 to À0.2 V (vs. Ag/AgCl
3m KCl) of 5 deposited onto a glassy carbon (GC) electrode in
0.1m KCl/water reveal a chemically reversible redox couple
with a half-wave potential E1/2 of À0.70 V (vs. Ag/AgCl 3m KCl,
Figure 1a, c). Compared with the previously known ÀNH2-
modified complex (À0.57 V vs. Ag/AgCl 3m KCl, adsorbed on
a GC electrode in 0.1m phosphate buffer solution), this value is
shifted to even more negative potentials (DE=À130 mV) as
expected from the stronger electron-donating effect of the
bpy-bound N,N’-dimethylamino groups. Voltammograms re-
corded with a bare electrode in the same potential range
show no redox signals (Figure 1a, a). Voltammetric charac-
terization of 5 in solution was conducted in 0.15m NBu4BF4/
MeCN at a Pt electrode. The I/E curve (not shown) of 5 shows
a chemically reversible redox couple with an estimated half-
wave potential of E1/2 =À0.63 V (vs. Ag/Ag+ 0.01m AgClO4).
This value is shifted to higher potentials compared with the
value obtained in aqueous electrolytes. However, the use of
different reference electrodes prevents a direct comparison of
the measured redox potentials.
To retain the redox potential of the Os-complex as low as
possible we exchanged the 2-bromopyridine precursor in our
earlier published synthesis[33] by DMAP, which contains a
ÀNMe2 group in para position to the coordinating N-atom.
Cross-coupling of stannate 2 (synthesized from DMAP in two
steps through 3; see the Supporting Information) and 6 gives
bipyridine 7 in good yield (91%). Deprotection under acidic
conditions leads to the carboxaldehyde 8. Reductive amination
with mono-protected ethylene diamine 9[38] gives the bipyri-
dine derivative 10. Removal of the Boc-protecting group was
successfully achieved with HCl in dioxane to give ligand 11.
The ethylene diamine linker in 11 allows for the covalent at-
tachment of the Os complex to the polymer backbone by reac-
tion of either the secondary or primary amino group with the
epoxy functions at the polymer chain.
Exchange of the chloro ligands in 5 by 11 in ethylene glycol
at 1208C, followed by treatment with NH4PF6/water yields
target complex 12 as the hexafluorophosphate salt. Cyclic vol-
tammetry of 12 in in 0.15m NBu4PF6/MeCN at a Pt disk elec-
trode (Figure 1b) reveals a chemically reversible redox couple
with E1/2 = +0.13 V (vs. Ag/Ag+ 0.01m AgClO4), the differential
pulse voltammogram of 12 (Figure S1, the Supporting Informa-
tion) in the same electrolyte (note that due to the low solubili-
ty of 12 in water, characterization in aqueous electrolyte was
not conducted) recorded with a pulse amplitude of 15 mV re-
veals a potential of maximum current (Emax) at +100 mV (vs.
Ag/Ag+ 0.01m AgClO4). The attachment of the ÀNMe2 to the
bpy-ligands leads to the envisaged low redox potential of the
redox mediator.
Recently, we described the synthesis of an amino-modified
bpy-ligand (bpyamino) suitable for the attachment to an epoxy-
modified polymer backbone by a ring-opening reaction be-
tween the nucleophilic ÀNH2 and the anchoring group.[33] This
ligand was synthesized starting from 2-bromopyridine and
acetal 6 (Scheme 2; for the synthesis of 6 see Scheme S1 in the
Supporting Information and ref. [30]). Ligand exchange with
[Os(bpy)2Cl2] leads to the complex [Os(bpy)2(bpyamino)]2+. The
complex showed high stability and was used as internal poten-
tial standard in a miniaturized pH sensor over a wide pH
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5321
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Synthesis of the Os complex-modified redox polymer P-12
rived to be +32.5 mV (with Epulse =pulse amplitude in the DPV
in mV). Note, that this value is shifted by about 70 mV to more
cathodic potentials than that obtained for the freely diffusing
complex 12 in the MeCN-based electrolyte (+105 mV vs. Ag/
Ag+ 0.01m AgClO4). This shift is identical to the potential dif-
ference between the aqueous electrolyte and the MeCN-based
electrolyte observed for 5 (aqueous electrolyte: À0.70 V vs.
Ag/AgCl 3m KCl; MeCN-based electrolyte: À0.63 V vs. Ag/Ag+
0.01m AgClO4). The half-wave potential of P-12 drop-cast onto
a graphite electrode was estimated to be +20 mV (the Sup-
porting Information, Figure S2). The applied synthetic strategy
leads to a redox polymer with the envisaged redox potential
close to 0 V versus Ag/AgCl 3m KCl.
The copolymer backbone (terpolymer P) for the modification
with the Os complex was synthesized (Scheme 3) from three
monomers, that is, styrene, butyl acrylate, and allyl methacry-
late, in a free radical polymerization reaction initiated by
azobisisobutyronitrile (AIBN). In a post-polymerization reaction,
Enzyme electrodes and biocatalytic activity
To characterize the redox polymer P-12 as immobilization
matrix for the wiring of enzymes to electrode surfaces, graph-
ite electrodes were modified with pyrroloquinoline quinone-
dependent soluble glucose dehydrogenase (PQQ-sGDH) and P-
12 in the presence of 2,2’-(ethylenedioxy)diethanethiol acting
as a crosslinker. The dithiol crosslinker ensures that stable films
are formed on the electrode surface by reaction of the ÀSH
groups with remaining epoxy functions of different polymer
chains.[30] In the following, the enzyme is designated as PQQ-
sGDH, whereas the enzyme integrated within the redox hydro-
gel is designated as PQQ-sGDH/P-12. PQQ-sGDH was chosen
because it is insensitive to oxygen,[39,40] which is crucial for bio-
sensors that operate under ambient conditions and for
oxygen/glucose enzymatic biofuel cells. PQQ-sGDH shows an
efficient mediated electron transfer in combination with Os
complex-modified redox hydrogels.[30] The redox potential of
Scheme 3. Synthesis of redox polymer P-12. The terpolymer backbone (P)
was obtained through a free radical polymerization reaction with AIBN as in-
itiator (nominal composition of P: k=50 mol%, l=35 mol%, m=15 mol%).
Epoxidation of the allylic double bonds with dimethyldioxirane (DMDO)
allows for a covalent attachment of the amino-modified Os complex 12.
Note that for P-12 only the molecular structure of the product is shown,
which is formed through the reaction of the primary amine in 12 and
a single epoxy group in P. The nominal concentration of the Os complex
within the modified polymer is expected to be <15 mol% because of non-
quantitative conversion in each reaction step and potential multiple binding
of the primary amino group to epoxy functions within the polymer.
PQQ-sGDH was reported to be À194 mV (vs. Ag/AgCl 3.5m
[14,41]
KCl, in the presence of Ca2+
)
However, it depends strongly
on the electrode material, pH value, presence of bivalent cat-
ions and/or presence of a redox mediator (see ref. [14] and ref-
erences therein). As predicted, the redox potential of PQQ-
sGDH is below that of P-12 (ꢀ +30 mV vs. Ag/AgCl 3m KCl)
and hence a sufficient thermodynamic driving force exists for
fast electron transfer from the enzyme to the electrode by
electron hopping along the redox hydrogel.
the allyl groups were transformed into epoxides under mild
conditions by means of dimethyl dioxirane (DMDO).[19] The
epoxy groups allow for the covalent attachment of the Os
complex 12 through a ring-opening reaction with the amino
functions in 12 in a MeOH/isopropyl alcohol mixture at elevat-
ed temperatures. In the following the resulting polymer will be
designated as P-12. Note that the binding of the complex may
occur through the primary and the secondary amine in 12. For
clarity reasons, only the product from the reaction of the pri-
mary amine and one epoxy function is shown in Scheme 3.
The redox activity of the modified polymer P-12 after drop-
casting on an electrode surface was evaluated by means of
cyclic voltammetry and differential pulse voltammetry (DPV) in
aqueous electrolytes. Cyclic voltammograms (Figure 1c) of
graphite electrodes coated with P-12 by drop-casting show
a chemically reversible redox couple with a half-wave potential
of around +30 mV (scan rate=100 mVsÀ1, PBS buffer, pH 7.4).
The DPV of P-12 in 3m KCl/water deposited onto a Pt elec-
trode (Figure 1d) shows well-defined and pronounced peak
with Emax = +25 mV (vs. Ag/AgCl 3m KCl). From this value,
a half-wave potential according to E1/2 =Emax +(Epulse/2) was de-
Figure 2a shows the current response of PQQ-sGDH/P-12-
modified electrodes prepared with different enzyme-to-poly-
~
&
*
^
mer mass ratios of 1:1 ( ), 1:2 ( ), 1:4 ( ), 1:5 ( ) in the pres-
ence of increasing glucose concentrations and a constant ap-
plied potential of +150 mV versus Ag/AgCl 3m KCl in 4-(2-hy-
droxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer
(10 mm, pH 7, 0.1 vol.% Triton X-100). The PQQ-sGDH/P-12-
modified electrodes show biocatalytic activity at comparatively
low operating potential of +150 mV versus Ag/AgCl 3m KCl
(see the Supporting Information, Figure S3 for cyclic voltam-
mograms of PQQ-sGDH/P-12-modified electrodes in presence
and absence of glucose). Michaelis–Menten-type calibration
graphs with apparent Km values of 0.98 mm to 2.98 mm are ob-
tained. The highest catalytic currents were obtained with a 1:4
enzyme-to-polymer ratio. In addition, the effect of different
crosslinker loadings was studied. Figure 2b shows glucose cali-
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5322
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
bration curves of PQQ-sGDH/P-12 films (1:4) with different
amounts of crosslinker used during drop-casting. The highest
currents were obtained for a crosslinker-to-polymer mass ratio
of ꢀ1:3.4. Interestingly, an increase of the temperature from
room temperature to 308C and to 378C results in a decrease
of the catalytic current (see Figure S4 in the Supporting Infor-
mation), most likely due to desorption of the PQQ-sGDH/P-12
films. Moreover, Sode et al. observed that the activity of PQQ-
sGDH decreases at temperatures above 378C to 65%.[42] Disso-
lution or desorption of the film in combination with a reduced
activity at elevated temperature may lead to the low catalytic
currents obtained at 378C. Further experiments were per-
formed at room temperature and with an enzyme-to-polymer
mass ratio of 1:4 and a crosslinker-to-polymer mass ratio of
1:3.4.
The calibration curve (Figure 2c) for a PQQ-sGDH/P-12 elec-
trode in phosphate-buffered saline (PBS) with a pH 7.4 (1 mm
CaCl2) shows again a Michaelis–Menten-type behavior with an
apparent KM value of 1.74 mm. A limiting current is reached at
concentrations above 5 mm. This behavior is in line with the
results obtained in HEPES buffer (Figure 2a,b). A linear current
response for low glucose concentration (0 to 5 mm) is, howev-
er, absent (inset in Figure 2c).
To evaluate the effect of parasitic oxygen reduction that was
reported for Os complexes with redox potentials < +70 mV
(vs. Ag/AgCl/3m NaCl),[32] chronoamperometry experiments in
HEPES buffer (10 mm, pH 7, 0.1 vol.% Triton X-100) and with
a glucose concentration of 20 mm (saturation concentration) in
the presence (air and oxygen atmosphere) and absence (argon
atmosphere) of O2 were conducted. Figure 3a shows the cata-
lytic currents derived in Ar- (left column), O2- (middle column),
and air-saturated electrolyte solutions (right column) at an ap-
plied potential of +150 mV (vs. Ag/AgCl 3m KCl). The current
values in Ar- and air-saturated electrolytes reveal almost identi-
cal values of about 280 nA. The catalytic current is slightly
higher for the O2-saturated solution; however, the differences
are small. Since parasitic oxygen reduction at the Os center
should lead to decreased current values, we conclude that cat-
alytic reduction at the suggested Os complex-modified poly-
mer P-12 is negligible.
The use of amperometric biosensors under physiological
conditions is often hampered by interferences that occur be-
cause of the oxidation of electroactive substances at high op-
erating potentials. Figure 3b shows the normalized catalytic
currents at a glucose concentration of 30 mm in the absence
and presence of electroactive interfering substances, that is,
uric acid and ascorbic acid. The deviation of the normalized
current, that is, the catalytic current in the presence of glucose
and the interfering substance divided by the steady-state
current that is obtained for only glucose was determined
to be 1.03 for glucose/uric acid and 1.13 glucose/ascorbic
acid. These results demonstrate the advantage of the low
redox potential of the designed Os complex-modified redox
hydrogel.
Figure 2. Biocatalytic activity of PQQ-sGDH/P-12 films (on graphite electro-
des, 3.05 mm) in dependence on the glucose concentration at an applied
potential of +150 mV versus Ag/AgCl 3m KCl; a) dependence of the catalyt-
ic current of PQQ-sGDH/P-12 films prepared with different enzyme-to-poly-
~
&
*
^
mer weight ratios 1:1 ( ), 1:2 ( ), 1:4 ( ), 1:5 ( ) in HEPES buffer (10 mm,
pH 7, 0.1 vol.% Triton X-100); b) dependence of the catalytic current of
PQQ-sGDH/P-12 films prepared with different crosslinker loadings and con-
&
*
^
stant enzyme-to-polymer ratio (1:4): 10.6 mg ( ), 21.3 mg ( ), 31.9 mg ( ) in
HEPES buffer (10 mm, pH 7, 0.1 vol% Triton X-100); c) calibration curve of
a PQQ-sGDH/P-12 film in PBS buffer (pH 7.4, 1 mm CaCl2); inset shows the
current response in a concentration range of 0 to 5 mm. Lines are guidance
for eyes only.
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5323
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Effect of oxygen (a), uric acid and ascorbic acid (b) on the catalytic
currents of PQQ-sGDH/P-12-modified graphite electrodes (3.05 mm) at glu-
cose saturation concentration of 30 mm and an applied potential of
+150 mV versus Ag/AgCl 3m in HEPES buffer (10 mm, pH 7, 0.1 vol% Triton
X-100); a) catalytic currents derived in different buffer solutions saturated
with argon, oxygen and air; error bars are standard deviations from at least
3 independent measurements; b) Normalized catalytic currents derived for
pristine glucose (=1) and in the presence of uric acid and ascorbic acid. Nor-
malized currents were calculated by dividing the absolute steady state cur-
rents obtained in the presence of glucose and the interfering species by the
steady state current that is obtained for only glucose.
&
Figure 4. Biofuel cell characteristics of PQQ-sGDH/P-12 ( ), FAD-rGDH/P-12
&
*
(
), and FAD-CtCDH/P-12 ( )-modified graphite electrodes (3.05 mm). a) Cur-
rent densities of the different P-12-based bioanodes in PBS (12 mm, pH 7.4,
0.14m NaCl, 2.7 mm KCl); applied potential +150 mV (vs. Ag/AgCl/3m KCl);
b) dependence of the power densities on the operating voltage at a glucose
concentration of 5 mm in PBS (12 mm, pH 7.4, 0.14m NaCl, 2.7 mm KCl, for
the experiments with PQQ-sGDH CaCl2 (1 mm) was added to stabilize PQQ-
sGDH) of the different bioanodes in combination with a MvBOx-based bioca-
thode (6 mm); all electrodes were prepared by drop-casting on graphite
electrodes.
Anodes for enzymatic biofuel cells
brane-less biofuel cell together with a bilirubin oxidase (from
Myrothecium verrucaria, MvBOx)-based biocathode in PBS
buffer solution at pH 7.4 containing glucose (5 mm). To ensure
that the bioanode is limiting the EBFC, the area of the MvBOx-
modified cathode was substantially larger than the area of the
bioanode. Figure 4b shows the dependence of the power den-
sity at different cell operating voltages. The highest power
density is again achieved with the PQQ-sGDH/P-12 couple. The
EBFCs show open-circuit voltages (OCVs) between 0.50 V and
0.54 V as expected from the redox potentials of the P-12 poly-
mer (ꢀ +0.02 V at graphite electrodes) and that of the MvBOx
cathode (around +0.45 V).
Redox hydrogels with low formal potentials are of particular in-
terest for enzymatic biofuel cells (EBFC) because the difference
between the redox potential of the bioanode and biocathode
determines the open circuit voltage of the resulting EBFC and
hence should be maximal. Consequently, the application of P-
12 in combination with other glucose-dependent and O2 toler-
ant enzymes, that is, flavin adenine dinucleotide-dependent re-
combinant glucose dehydrogenase (FAD-rGDH), and the flavo-
dehydrogenase domain of cellubiose dehydrogenase from Cor-
ynascus thermophilus (FAD-CtCDH), as well as PQQ-sGDH as po-
tential anodes in biofuel cells was investigated.
Figure 4a shows the current density of graphite electrodes
modified with films of PQQ-sGDH/P-12, FAD-rGDH/P-12, and
FAD-CtCDH/P-12 at different glucose concentrations (0–
50 mm) obtained at an applied potential of +150 mV (vs. Ag/
AgCl 3m KCl) in PBS buffer (pH 7.4). The highest current densi-
ty is obtained with PQQ-sGDH followed by FAD-rGDH and
FAD-CtCDH.
Conclusion
An Os complex was designed envisaging stable N6 coordina-
tion, low redox potential due to stable electron-donating sub-
stituents at bpy ligands, and an amino-terminated linker chain
for binding to epoxy-functionalities of a terpolymer backbone.
A suitable Os complex was synthesized in a multi-step ap-
proach starting from simple precursors, that is, N,N’-dimethyl-
Electrodes modified with PQQ-sGDH/P-12, FAD-rGDH/P-12,
and FAD-CtCDH/P-12 films were used as bioanodes in a mem-
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5324
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Roche Diagnostics (Penzberg). Pyrroloquinolinequinone (PQQ) was
purchased from Fluka. The apo-enzyme sGDH (1.08 mg) was recon-
stituted in 10 mm HEPES buffer (30 mL) containing 150 mm CaCl2
(150 mm) and PQQ (500 mm) to form the holoenzyme PQQ-sGDH
(incubation time 30 min, 48C). The activity of PQQ-sGDH was deter-
mined using a UV/Vis spectrophotometric assay containing 2,6-di-
chloroindophenol (DCIP). The absorption of DCIP was monitored at
aminopyridin derivatives. The introduction of the ÀNMe2 group
in the ligand sphere of the Os complex ensures the envisaged
low redox potential of around 0 V (vs. Ag/AgCl 3m KCl). The
covalent attachment of the Os complexes to the hydrogel was
achieved by epoxidation of the polymer backbone and the re-
action of the formed epoxy functions with amino groups
within the ligand system of the Os species. The redox hydrogel
could be successfully used as immobilization matrix for electri-
cal wiring of various enzymes (PQQ-sGDH, FAD-rGDH, FAD-
CtCDH) to graphite electrodes. The modified electrodes were
characterized with respect to their potential use as interfer-
ence-free amperometric glucose sensors and as anodes in
membrane-less glucose/oxygen enzymatic biofuel cells in com-
bination with a bilirubin oxidase-based biocathode.
a
wavelength of 600 nm. The activity of PQQ-sGDH was
3827 UmgÀ1
.
Flavin adenine dinucleotide (FAD)-dependent recombinant glucose
dehydrogenase (rGDH) was prepared from Glomerella cingulata
and was recombinantly expressed in Pichia pastoris (FAD-rGDH) as
reported previously.[48] The protein concentration was 21 mgmLÀ1
,
the specific activity of the enzyme was 793 UmgÀ1
.
The flavodehydrogenase domain of cellubiose dehydrogenase
(CDH) from Corynascus thermophilus (FAD-CtCDH) was prepared by
recombinant production in Pichia pastoris.[49] The protein concen-
Evidently, the carefully designed low overpotential of the
proposed redox polymer with respect of the PQQ-moiety in
the active site of the enzyme is leading to a lower driving
force for the electron transfer reaction between the enzyme
and the polymer-bound redox mediator. Although higher cur-
rent densities have been obtained previously with redox poly-
mers bearing redox mediators exhibiting a higher formal po-
tential, these higher current densities were obtained at the ex-
pense of a decrease of the open-circuit voltage of a related
biofuel cell. In the case of a sensor application, a higher media-
tor potential is leading to cooxidation of potentially interfering
compounds. Thus, although the power density and the biosen-
sor activity/sensitivity of the proposed electrode architecture
are lower as compared with those reported for other PQQ-
sGDH-based enzyme electrodes that operate in a direct elec-
tron transfer mode[40,43] or in combination with different media-
tor species,[14,32,37,44–46] our results clearly demonstrate the ad-
vantage of an optimized redox potential preventing unintend-
ed electrochemical side reactions. The optimization/adjustment
of the formal potential of the mediator with respect to a given
enzyme and possible interferences is crucial for the design of
improved enzyme-based biosensors and biofuel cells.
tration and the specific activity were 16 mgmLÀ1 and 268 UmgÀ1
respectively.
,
Bilirubin oxidase from Myrothecium verrucaria (MvBOx) was ob-
tained from Sigma–Aldrich. For electrode preparations, the enzyme
(20 mgmLÀ1) was dissolved in phosphate buffer (100 mm, pH 7)
and drop-cast on the graphite surface.
Graphite electrodes were prepared from graphite rods (diameter
3.05 mm; SGL Carbon). The electrodes were polished with emery
paper and subsequently sonicated for 5 min in water and in etha-
nol, respectively. The crosslinker, 2,2’-(ethylenedioxy)diethanethiol,
was diluted in ethanol (1:50 v/v) prior to use. The redox polymer
P-12 (5 wt% in DMSO) was drop-cast on the electrode surface by
means of a pipette (polymer amount 72 mg, 1.44 mL polymer sus-
pension) and the electrode was left to dry in air. Then, enzyme
(18 mg, dissolved in HEPES buffer, 10 mm, pH 7, 0.1 vol.% Triton X-
100) and crosslinker (21.3 mg) were successively dropped on top of
the polymer layer and mixed thoroughly on the electrode surface
by means of a pipette tip. The modified electrodes were kept at
48C overnight to complete the crosslinking process. The modified
electrodes were gently rinsed with water to remove loosely bound
polymer and enzyme as well as remaining crosslinker. For biofuel
cell tests, the biocathode was constructed by absorbing MvBOx
(800 mg) on a 6 mm graphite electrodes using a drop-cast process.
The electrode was stored at 48C.
NMR experiments were conducted with a DPX 200 or a DRX 400
spectrometer from Bruker with 1H resonance frequencies of
200.13 MHz and 400.13 MHz, respectively. UV/Vis absorption spec-
tra were recorded with a Cary 60 absorption spectrometer from
Agilent in quartz cuvettes (optical path length=1 cm) in MeOH.
MS spectra were recorded with VG Instruments Autospec (EI), Fin-
nigan GCQ Iontrap (FAB) or a LTQ-Orbitrap XL spectrometer (ESI).
Experimental Section
Materials and methods
All materials and chemicals were purchased from Sigma–Aldrich,
J.T. Baker, Acros Organics, VWR chemicals, Alfa Aesar or Applichem.
Detailed protocols for the syntheses as well as characteristic analyt-
ical data for all compounds are given in the Supporting Informa-
tion. The synthesis of 4,4’-bis-(N,N-dimethylamino)-2,2’-bipyridine
was based on protocols described in ref. [36]. The synthesis of the
amino-modified bipyridine ligand 11 was performed following pro-
tocols described in ref. [33] The synthesis of the copolymer back-
bone and the modification with the Os complex was based on pro-
cedures reported in ref. [19] Dimethyldioxirane (DMDO) was freshly
prepared prior to use according to ref. [47].
Electrochemical measurement
All electrochemical measurements were performed in a convention-
al three electrode set-up comprised of a graphite (diameter=
0.305 cm/surface area=0.073 cm2), platinum (1 mm/0.79 mm2) or
glassy carbon (0.3 cm/0.071 cm2) working electrode, an Ag/AgCl
3m KCl reference electrode and a Pt wire acting as counter elec-
trode. Chronoamperometry was performed with a CHI1030 eight-
channel potentiostat (CH Instruments) at an applied potential of
+150 mV. All buffer solutions were purged with Ar for 15 min prior
to the measurements. Differential pulse voltammetry (DPV) and
cyclic voltammetry (CV) were used to investigate the redox poten-
tial of the redox polymer P-12 using a three-electrode setup and
potentiostats from PalmSens or CH Instruments. For experimental
4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer
(10 mm, pH 7.0) containing 0.1 vol.% Triton X-100 was prepared by
dissolving HEPES in deionized water (Millipore) and adjustment of
the pH value by addition of solid NaOH. Phosphate-buffered saline
(PBS, 12 mm, pH 7.4) was prepared with Millipore water containing
NaCl (0.14m) and KCl (2.7 mm). The apo-enzyme of PQQ-depen-
dent soluble glucose dehydrogenase (PQQ-sGDH) was a gift from
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5325
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[11] P. Ó Conghaile, S. Pçller, D. MacAodha, W. Schuhmann, D. Leech, Bio-
sens. Bioelectron. 2013, 43, 30–37.
[12] T. J. Ohara, R. Rajagopalan, A. Heller, Anal. Chem. 1993, 65, 3512–3517.
[13] D. A. Guschin, H. Shkil, W. Schuhmann, Anal. Bioanal. Chem. 2009, 395,
1693–1706.
[14] V. Flexer, N. Mano, Anal. Chem. 2014, 86, 2465–2473.
[15] D. A. Guschin, J. Castillo, N. Dimcheva, W. Schuhmann, Anal. Bioanal.
Chem. 2010, 398, 1661–1673.
[16] U. Salaj-Kosla, M. D. Scanlon, T. Baumeister, K. Zahma, R. Ludwig, P. Ó.
Conghaile, D. MacAodha, D. Leech, E. Magner, Anal. Bioanal. Chem.
2013, 405, 3823–3830.
[17] M. Shao, D. A. Guschin, Z. Kawah, Y. Beyl, L. Stoica, R. Ludwig, W. Schuh-
mann, X. Chen, Electrochim. Acta 2014, 128, 318–325.
[18] M. N. Zafar, F. Tasca, S. Boland, M. Kujawa, I. Patel, C. K. Peterbauer, D.
Leech, L. Gorton, Bioelectrochemistry 2010, 80, 38–42.
[19] S. Pçller, Y. Beyl, J. Vivekananthan, D. A. Guschin, W. Schuhmann, Bioe-
lectrochemistry 2012, 87, 178–184.
[20] H.-H. Kim, N. Mano, Y. Zhang, A. Heller, J. Electrochem. Soc. 2003, 150,
A209–A213.
[21] P. Kavanagh, D. Leech, Tetrahedron Lett. 2004, 45, 121–123.
[22] H. Masui, A. B. P. Lever, Inorg. Chem. 1993, 32, 2199–2201.
[23] N. J. Ronkainen, H. B. Halsall, W. R. Heineman, Chem. Soc. Rev. 2010, 39,
1747–1763.
details see the main text. Measurement parameters for DPV experi-
ments: step potential=5 mV; modulation amplitude=15 mV;
modulation time=0.06 s and interval time 0.2 s. Cyclic voltammo-
grams were recorded with a step potential of 1 or 2 mV at different
scan rates (see main text). Half-wave potentials (E1/2) determined
from cyclic voltammetry were calculated from the peak potential
in the forward (Epf) and backward (Epb) scan of the corresponding
redox couples as E1/2 =(Epf +Epb)/2. For cyclic voltammetry of 5, the
dichloro complex was suspended in EtOH and drop-cast onto the
electrode surface (10–20 mL). The solvent was removed with
a gentle argon stream.
Biofuel cell measurement
Biofuel cell tests were conducted in a one-compartment electro-
chemical cell using an Autolab PGSTAT12 potentiostat. The anode
was connected as working electrode, whereas the cathode was
connected as combined counter and reference electrode. To
ensure that the anode is the limiting electrode in the biofuel cell,
the geometric surface area of the cathode (0.28 cm2) was always
larger than that of the anode (0.073 cm2).
[24] A. Heller, B. Feldman, Chem. Rev. 2008, 108, 2482–2505.
[25] S. Calabrese Barton, J. Gallaway, P. Atanassov, Chem. Rev. 2004, 104,
4867–4886.
[26] Z. Gao, G. Binyamin, H.-H. Kim, S. C. Barton, Y. Zhang, A. Heller, Angew.
Chem. Int. Ed. 2002, 41, 810; Angew. Chem. 2002, 114, 838.
[27] Y. Zhang, C. Ji, Anal. Chem. 2010, 82, 5275–5281.
[28] V. Soukharev, N. Mano, A. Heller, J. Am. Chem. Soc. 2004, 126, 8368–
8369.
Current and power density measurements were performed in PBS
buffer (12 mm, pH 7.4, 0.14m NaCl, 2.7 mm KCl) with a glucose
concentration of 5 mm. In case of the PQQ-sGDH/P-12 films CaCl2
(1 mm) was added to the buffer solution to stabilize the PQQ-
sGDH complex. For determination of the power output curves po-
tential pulses in the range from 0 to 0.6 V (against open-circuit
voltage) were applied and the current at each potential interval
was recorded when a steady state was reached.
[29] F. Mao, N. Mano, A. Heller, J. Am. Chem. Soc. 2003, 125, 4951–4957.
[30] P. Pinyou, S. Pçller, X. Chen, W. Schuhmann, Electroanalysis 2015, 27,
200–208.
[31] W. D. Harman, D. P. Fairlie, H. Taube, J. Am. Chem. Soc. 1986, 108, 8223–
8227.
[32] A. PrØvoteau, N. Mano, Electrochim. Acta 2012, 68, 128–133.
[33] S. Pçller, W. Schuhmann, Electrochim. Acta 2014, 140, 101–107.
[34] F. Barri›re, P. Kavanagh, D. Leech, Electrochim. Acta 2006, 51, 5187–
5192.
[35] S. Murata, M. Miura, M. Nomura, J. Chem. Soc. Chem. Commun. 1989,
116.
[36] D. Cuperly, P. Gros, Y. Fort, J. Org. Chem. 2002, 67, 238–241.
[37] K. Habermüller, A. Ramanavicius, V. Laurinavicius, W. Schuhmann, Elec-
troanalysis 2000, 12, 1383–1389.
Acknowledgements
Financial support by the European Commission in the frame-
work of the ITN “Bioenergy” (PITN-GA-2013-607793) and by the
Deutsche Forschungsgemeinschaft in the framework of the
Cluster of Excellence RESOLV (EXC 1069) is acknowledged. P.P.
thanks the Deutscher Akademischer Austauschdienst (DAAD)
for a PhD scholarship. The authors thank D. Wolters for ESI
measurements.
[38] M. A. Ghanem, J.-M. ChrØtien, A. Pinczewska, J. D. Kilburn, P. N. Bartlett,
J. Mater. Chem. 2008, 18, 4917–4927.
[39] L. Ye, M. Haemmerle, A. J. J. Olsthoorn, W. Schuhmann, H.-L. Schmidt,
J. A. Duine, A. Heller, Anal. Chem. 1993, 65, 238–241.
[40] C. Tanne, G. Gçbel, F. Lisdat, Biosens. Bioelectron. 2010, 26, 530–535.
[41] A. Sato, K. Takagi, K. Kano, N. Kato, J. A. Duine, T. Ikeda, Biochem. J.
2001, 357, 893–898.
[42] K. Sode, K. Watanabe, S. Ito, K. Matsumura, T. Kikuchi, FEBS Lett. 1995,
364, 325–327.
[43] I. W. Schubart, G. Gçbel, F. Lisdat, Electrochim. Acta 2012, 82, 224–232.
[44] A. Malinauskas, J. Kuzmarskyt, R. Meskys, A. Ramanavicius, Sens. Actua-
tors B 2004, 100, 387–394.
Keywords: biosensors · hydrogel · osmium · polymers · redox
chemistry
[1] A. Heller, Curr. Opin. Chem. Biol. 2006, 10, 664–672.
[2] K. Sirkar, M. V. Pishko, Anal. Chem. 1998, 70, 2888–2894.
[3] T. J. Ohara, R. Rajagopalan, A. Heller, Anal. Chem. 1994, 66, 2451–2457.
[4] D. A. Guschin, Y. M. Sultanov, N. F. Sharif-Zade, E. H. Aliyev, A. A. Efen-
diev, W. Schuhmann, Electrochim. Acta 2006, 51, 5137–5142.
[5] E. Suraniti, S. Tsujimura, F. Durand, N. Mano, Electrochem. Commun.
2013, 26, 41–44.
ˇ
˙
ˇ
ˇ
ˇ
[45] J. Razumiene˙, V. Gureviciene, V. Laurinavicius, J. V. Grazulevicius, Sens.
Actuators B 2001, 78, 243–248.
[6] F. Tasca, L. Gorton, M. Kujawa, I. Patel, W. Harreither, C. K. Peterbauer, R.
Ludwig, G. Nçll, Biosens. Bioelectron. 2010, 25, 1710–1716.
[7] M. Shao, M. N. Zafar, M. Falk, R. Ludwig, C. Sygmund, C. K. Peterbauer,
D. A. Guschin, D. MacAodha, P. Conghaile, D. Leech, M. D. Toscano, S.
Shleev, W. Schuhmann, L. Gorton, ChemPhysChem 2013, 14, 2260–2269.
[8] T. Kothe, N. PlumerØ, A. Badura, M. M. Nowaczyk, D. A. Guschin, M.
Rçgner, W. Schuhmann, Angew. Chem. Int. Ed. 2013, 52, 14233–14236;
Angew. Chem. 2013, 125, 14483–14486.
[46] P. Pinyou, F. Conzuelo, K. Sliozberg, J. Vivekananthan, A. Contin, S.
Pçller, N. PlumerØ, W. Schuhmann, Bioelectrochemistry 2015, 106, 22–
27.
[47] W. Adam, J. Bialas, L. Hadjiarapoglou, Chem. Ber. 1991, 124, 2377.
´
[48] C. Sygmund, P. Staudigl, M. Klausberger, N. Pinotsis, K. Djinovic-Carugo,
L. Gorton, D. Haltrich, R. Ludwig, Microb. Cell Fact. 2011, 10, 106.
[49] W. Harreither, A. K. G. Felice, R. Paukner, L. Gorton, R. Ludwig, C. Syg-
mund, Biotechnol. J. 2012, 7, 1359–1366.
[9] D. R. Baker, R. F. Simmerman, J. J. Sumner, B. D. Bruce, C. A. Lundgren,
Langmuir 2014, 30, 13650–13655.
[10] A. Badura, D. Guschin, T. Kothe, M. J. Kopczak, W. Schuhmann, M.
Rçgner, Energy Environ. Sci. 2011, 4, 2435.
Received: November 15, 2015
Published online on March 1, 2016
Chem. Eur. J. 2016, 22, 5319 – 5326
www.chemeurj.org
5326
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim