An oxygen cathode operating in a physiological solution

Article abstract of DOI:10.1021/ja025874v

We report the electroreduction of O₂ to water under physiological conditions (pH 7.4, 0.15 M NaCl, 37.5 °C) at a current density of 5 mA cm^-2 and at a potential only 0.18 V reducing versus that of the reversible O₂/H₂

Full text of DOI:10.1021/ja025874v

Published on Web 05/10/2002
An Oxygen Cathode Operating in a Physiological Solution
Nicolas Mano, Hyug-Han Kim, Yongchao Zhang, and Adam Heller*
Contribution from the Department of Chemical Engineering and the Texas Materials Institute,
The UniVersity of Texas, Austin, Texas 78712
Received February 8, 2002
2
Abstract: We report the electroreduction of O to water under physiological conditions (pH 7.4, 0.15 M
-
2
NaCl, 37.5 °C) at a current density of 5 mA cm and at a potential only 0.18 V reducing versus that of the
reversible O /H O electrode at pH 7.4. The immobilized electrocatalyst enabling the reduction is the
2
2
electrostatic adduct of bilirubin oxidase from Myrothecium verrucaria, a polyanion at pH >4.1, and the
polycationic redox copolymer of polyacrylamide and poly (N-vinylimidazole) complexed with [Os (4,4′-
+
/2+
dichloro-2,2′-bipyridine)
2
Cl]
, cross-linked on carbon cloth. The current density of the rotating electrodes
-
2
-2
was O transport limited up to 8.8 mA cm ; their kinetic limit was reached at 9.1 mA cm . The operational
2
life of the electrodes depended on their angular velocity, which defined not only the current density but
also the mechanical shear stress stripping the electrocatalyst. When the electrodes were rotated at 300
rpm and were poised at -256 mV versus the potential of the reversible O
cm initial current density decreased to 1.3 mA cm after 6 days of continuous operation at 37.5 °C.
2
/H
2
O electrode, their 2.4 mA
-
2
-2
operated at pH 5 at a current density of 175 µA cm- when
2
Introduction
poised near the thermodynamic potential of the reversible O2/
The four-electron electroreduction of O2 to water under
physiological conditions is of interest because miniature biofuel
1
5
H2O electrode. Tarasevich et al. also reported operation of an
electrode made of a composite of high surface area porous
carbon and an unidentified laccase at pH 3.5 at -0.2 V relative
to the thermodynamic potential at a current density approaching
1
cells with O2 electroreducing cathodes and glucose electro-
2,3
oxidizing anodes may power in the future sensors and
actuators implanted in the body. They are likely to be smaller
than batteries, because unlike batteries they do not require a
case and a seal.
-
2
1
0 mA cm . They did not provide, however, enough detail to
1
6
allow others to reproduce their experiment. Two recent
1
7,18
studies
describe the electroreduction of O2 to water, in the
The blue, copper-containing oxidases, examples of which
include laccases, ascorbate oxidase, and bilirubin oxidase,
absence of chloride, in a pH 5 citrate buffer at a current density
-
2
of 5 mA cm at -0.13 V versus the reversible potential of the
O2/H2O electrode at 37.5 °C. The electrocatalyst of these
electrodes was a composite of laccase from Coriolus hirsutus
cross-linked with a redox polymer on a hydrophilic cloth of
4
-8
catalyze the four-electron reduction of O2 to water.
Among
these enzymes, the laccases were the most studied.9 Tarasevich
et al. reported that monolayers of laccases on vitreous carbon
catalyze the four-electron reduction of O2. The reported electrode
-14
1
0-µm-diameter carbon fibers. The redox polymer, PVI-
2
+/3+
Os(tpy)(dme-bpy)
of the imidazoles complexed with [Os(tpy)(dme-bpy)]2
, [poly-N-vinylimidazole with one-fifth
*
Author to whom correspondence should be addressed. E-mail: heller@
+/3+
che.utexas.edu.
(tpy
(
(
(
(
(
(
1) Chen, T.; Barton, S. C.; Binyamin, G.; Gao, Z.; Zhang, Y.; Kim, H.-H.;
)
terpyridine; dme-bpy ) 4,4′-dimethyl-2,2′-bipyridine)], elec-
Heller, A. J. Am. Chem. Soc. 2001, 123, 8630-8631.
2) Katz, E.; Willner, I.; Kotlyar, A. B. J. Electroanal. Chem. 1999, 479, 64-
trically connected (“wired”) the laccase reaction centers to the
fibers. A glucose/O2 biofuel cell built with the “wired” C.
hirsutus laccase cathode had a 5-fold higher power density than
6
8.
3) Katz, E.; B u¨ ckmann, A. F.; Willner, I. J. Am. Chem. Soc. 2001, 123, 10752-
0753.
4) Messerschmidt, A. Multi-Copper Oxidases; World Scientific: Singapore,
1
1
earlier glucose/O2 cells.
1
997.
5) Spiro, T. G. Copper Proteins; John Wiley and Sons: New York, 1981;
Vol. 3.
However, the current density of the “wired” C. hirsutus
laccase-carbon cloth cathode was very small at the physiologi-
cal pH of 7.4 and at the physiological chloride concentration
of 0.14 M, because the C. hirsutus laccase is practically inactive
6) Solomon, E. I.; Sundaram, U. M.; Machonkin, T. E. Chem. ReV. 1996, 96,
2
563-2605.
(
7) Solomon, E. I.; Lowery, M. D. Science 1993, 259, 1575-1581.
(
8) Palmore, G. T.; Whitesides, G. M. ACS Symp. Ser. 1994, No. 566, 271-
2
90.
-
at neutral pH and is inhibited by Cl . The current density of a
(
9) Palmore, G. T.; Kim, H.-H. J. Electroanal. Chem. 1999, 464, 110-117.
(
10) Yaropolov, A. I.; Kharybin, A. N.; Emmeus, J.; Marko-Varga, G.; Gorton,
L. Bioelectrochem. Bioenerg. 1996, 40, 49-57.
(15) Tarasevich, M. R.; Yaropolov, A. I.; Bogdanovskaya, V. A.; Varfolomeev,
S. D. Bioelectrochem. Bioenerg. 1979, 6, 393-403.
(
11) Santucci, R.; Ferri, T.; Morpurgo, L.; Savini, I.; Avigliano, L. Biochem. J.
1
998, 332, 611-615.
(16) Tarasevich, M. R.; Bogdanovskaya, V. A.; Gavrilova, E. F.; Orlov, S. B.
J. J. Electroanal. Chem. Interfacial Electrochem. 1986, 206, 217-227.
(17) Barton, S. C.; Kim, H.-H.; Binyamin, G.; Zhang, Y.; Heller, A. J. Phys.
Chem. 2001, 105, 11917-11921.
(
(
12) Trudeau, F.; Daigle, F.; Leech, D. Anal. Chem. 1997, 69, 882-886.
13) Thuesen, M. H.; Farver, O.; Reinhammar, B.; Ulstrup, J. Acta Chem. Scand.
1
998, 52, 555-562.
(
14) Gelo-Pujic, M.; Kim, H.-H.; Butlin, N. G.; Palmore, G. T. Appl. EnViron.
(18) Barton, S. C.; Kim, H.-H.; Binyamin, G.; Zhang, Y.; Heller, A. J. Am.
Chem. Soc. 2001, 123, 5802-5803.
Microbiol. 1999, 65, 5515-5521.
6480
9
J. AM. CHEM. SOC. 2002, 124, 6480-6486
10.1021/ja025874v CCC: $22.00 © 2002 American Chemical Society

O
2
Cathode Operating in Physiological Solution
A R T I C L E S
smooth vitreous carbon electrode, made by “wiring” the laccase
-
2
from Pleurotus ostreatus, nevertheless reached ∼100 µA‚cm
at a potential 0.19 V reducing relative to the reversible potential
at pH 7 in phosphate-buffered 0.15 M NaCl.19 The projected
current density for this electrode, were it made as a carbon cloth
-
2
composite, would have been ∼0.3 mA cm .
Tsujimura et al. recently described29 a carbon felt cathode
coated with bilirubin oxidase from Myrothecium Verrucaria, an
enzyme used in the clinical assay2
0-23
of serum bilirubin. The
enzyme catalyzes the oxidation of bilirubin to biliverdin (eq 1)
1
bilirubin + / O f biliverdin + H O
(1)
2
2
2
Figure 1. Structure of the bilirubin oxidase “wiring” redox polycation
+
/2+
PAA-PVI-[Os(dcl-bpy)2Cl]
.
and then to a yet unidentified purple pigment.2
4-28
The electrode
poised at -0.17 V versus the reversible potential of the O2/
H2O electrode operated at 0.5 mA cm only for 2 h in a pH 7
phosphate-buffered 0.1 M KCl solution.29
salt), L-sodium ascorbate, 4-acetaminophen, NaIO
4
, NaCl, NaOH,
-
2
KCNS, KBr, and NaF were purchased from Sigma (St. Louis, MO).
Poly(ethylene glycol) (400) diglycidyl ether (PEGDGE) was purchased
from Polysciences Inc. (Warrington, PA). A fresh solution of BOD in
pH 7.4 20 mM phosphate buffer (PB) was prepared daily. The uric
Bilirubin oxidase (BOD) is a monomeric protein with a
molecular mass of 60 kDa. Parts of its sequence are homologous
3
0
with those of other multicopper oxidases, such as laccases,
acid was dissolved in dilute NaOH and then neutralized with dilute
ascorbate oxidase,31 and ceruloplasmine having a copper-
binding His-Cys-His33 triad. The copper ions of these enzymes
32
35
H
3
PO
4
to yield a 10 mM aqueous solution. The electrochemical
measurements were performed in PBS (pH 7.4 20 mM phosphate-
buffered 0.15 M NaCl) except in the experiments where the pH and
anion dependences of the steady-state electroreduction of O were
2
studied. In these, borate, citrate, acetate, phosphate, and Tris buffers
were employed. All solutions were made with deionized water that
was passed through a purification train (Sybron Chemicals Inc.,
Pittsburgh, PA). Carbon cloth (Toray TGPH-030) was received, as a
are classified into three types by their optical and magnetic
4
properties. Type I (blue) copper ions have a characteristic Cys
to Cu (II) charge-transfer band near 600 nm. The type I copper
center accepts electrons from the electron-donating substrate
of the enzyme and relays these to the O2 reduction site. The
latter is a trinuclear cluster, consisting of a type II copper ion
and a type III pair of cupric ions with a characteristic 330-nm
sample, from E-TEK (Somerset, NJ). Ultrapure O
purchased from Matheson (Austin, TX).
2
and argon were
shoulder.4 Shimizu et al.
,5
33,34
has shown that BOD is also a
Synthesis of the Redox Polymer PAA-PVI-[Os(4,4′-dichloro-
+/2+
multicopper oxidase, containing one type I, one type II, and
two type III copper ions.
2
2,2′-bipyridine) Cl] . 4,4′-Dinitro-2,2′-bipyridine N,N′-dioxide was
prepared as described.3 4,4′-dichloro-2,2′-bipyridine (dcl-bpy) was
6,37
synthesized from 4,4′-dinitro-2,2′-bipyridine N,N′-dioxide by modifying
Here we describe persistent four-electron electroreduction of
O2 to water under physiological conditions at previously
unattained current densities and potentials. The reaction is
catalyzed by the electrostatic adduct of BOD and a redox
polymer tailored to donate protons and mediate the transport
of electrons from the electrode to the enzyme.
the procedure of Maerker et al.3
6,38
2 2
Os(dcl-bpy) Cl was prepared as
follows: (NH
in a 1:2 molar ratio and refluxed under argon for 1 h (yield 85%).
The Os(dcl-bpy) Cl was then complexed with the 1:7 polyacrylamide-
4 2 6
) OsCl and ′′dcl-bpy were dissolved in ethylene glycol
3
9
2
2
poly(N-vinylimidazole) (PAA-PVI) copolymer and purified as de-
39
scribed. Figure 1 shows the structure and the stoichiometry of the
+
/2+
Experimental Section
PAA-PVI-[Os(4,4′-dichloro-2,2′-bipyridine)
2
Cl]
redox polymer.
Electrodes. The carbon cloth electrodes were made by the reported
three-step procedure.17 Their substrates were 3-mm-diameter vitreous
carbon electrodes mounted in Teflon sleeves. In the first step, the
substrates were polished with 0.05-µm Al O powder (Buehler, Lake
2 3
Bluff, IL) rinsed and sonicated for 10 min in ultrapure water. The
Chemicals and Materials. Bilirubin oxidase (EC 1.3.3.5) from M.
Verrucaria, catalase from bovine liver (EC 1.11.1.6), uric acid (sodium
(
(
(
(
(
19) Barton, S. C.; Pickard, M.; Vasquez-Duhalf, R.; Heller, A. Biosens.
Bioelectron. 2002, 00, 00-00.
20) Andreu, Y.; Galban, J.; De Marcos, S.; Castillo, J. R. Fresenius J. Anal.
Chem. 2000, 368, 516-521.
polishing step was repeated until no voltammetric features beyond water
21) Doumas, B. T.; Wu, T. W.; Poon, K. C. P.; Jendrzejczak, B. Clin. Chem.
-
1
oxidation were observed in a 50 mV s scan in PBS through the 0.2-
and 1-V range. The electrodes were then dried in an air stream. In the
second step, the 350-µm-thick carbon cloth (nominal 78% void fraction,
composed of 10-µm-diameter fibers) was cut into 4-mm-diameter disks.
These were cemented, using conductive carbon paint (SPI, West
Chester, PA), to the surface of the substrate electrodes. The substrate-
1
985, 31, 1677-1682.
22) Lavin, A.; Sung, C.; Klibanov, A. M.; Langer, R. Science 1985, 230, 543-
45.
23) Kosaka, A.; Yamamoto, C.; Morisita, C.; Nakane, K. Clin. Biochem. 1987,
0, 451-458.
5
2
(
(
(
(
(
24) Tanaka, N.; Murao, S. Agric. Biol. Chem. 1983, 47, 1627-1628.
25) Murao, S.; Tanaka, N. Agric. Biol. Chem. 1982, 46, 2031-2034.
26) Tanaka, N.; Murao, S. Agric. Biol. Chem. 1982, 46, 2499-2503.
27) Tanaka, N.; Murao, S. Agric. Biol. Chem. 1985, 49, 843-844.
28) Koikeda, S.; Ando, K.; Kaji, H.; Inoue, T.; Murao, S.; Takeuchi, K.;
Samejima, T. J. Biol. Chem 1993, 268, 18801-18809.
2
bound cloth was made hydrophilic by exposure to a 1 Torr O plasma
40
for 5 min. The rotating ring-disk electrodes (RRDE), with 3-mm-
(
29) Tsujimura, S.; Tatsumi, H.; Ogawa, J.; Shimizu, S.; Kano, K.; Ikeda, T. J.
Electroanal. Chem. 2001, 496, 69-75.
diameter vitreous carbon disks and platinum rings were similarly made.
In these, the platinum electrodes were cycled in 0.5 M H SO until the
2 4
(
30) Ducos, V.; Brzowski, A. M.; Wilson, K. S.; Brown, S. H.; Ostergaard, P.;
Schneider, P.; Yaver, D. S.; Pedersen, A. H.; Davies, G. J. Nat. Struct.
Biol. 1998, 5, 310-316.
(35) Binyamin, G.; Chen, T.; Heller, A. J. Electroanal. Chem. 2001, 500, 604-
611.
(
(
(
(
31) Messerschmidt, A.; Ladenstein, R.; Huber, R. J. Mol. Biol. 1992, 224, 179-
2
05.
(36) Anderson, S.; Constable, E. C.; Seddon, K. R.; Turp, E. T.; Baggott, J. E.;
Pilling, J. J. Chem. Soc., Dalton Trans. 1985, 2247-2250.
(37) Kenausis, G.; Taylor, C.; Rajagopalan, R.; Heller, A. J. Chem. Soc., Faraday
Trans. 1996, 92, 4131-4135.
32) Zaitseva, I.; Zaitsev, V.; Card, G.; Moshkov, K.; Bax, B.; Ralph, A.; Lindley,
P. J. Biol. Inorg. Chem. 1996, 1, 15-23.
33) Shimizu, A.; Kwon, J. H.; Sasaki, T.; Satoh, T.; Sakurai, N.; Sakurai, T.;
Yamaguchi, S.; Samejima, T. Biochemistry 1999, 38, 3034-3042.
34) Shimizu, A.; Sasaki, T.; Kwon, J. H.; Odaka, A.; Satoh, T.; Sakurai, N.;
Sakurai, T.; Yamaguchi, S.; Samejima, T. J. Biochem. 1999, 125, 662-
(38) Maerker, G.; Case, F. H. J. Am. Chem. Soc. 1958, 80, 2475-2477.
(39) Zakeeruddin, S. M.; D. M. Fraser, D. M.; Nazeeruddin, M.-K.; Gratzel,
M. J. Electroanal. Chem. 1992, 337, 253-256.
6
68.
(40) Sayka, A.; Eberhart, J. G. Solid State Technol. 1989, 32, 69-70.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002 6481

A R T I C L E S
Mano et al.
voltamogramms showed the characteristics of a clean platinum
electrode. In the third step, the electrocatalyst was deposited on the
carbon cloth. The deposition solution consisted of 10 µL of 10 mg
-
1
mL aqueous redox polymer solution, 2 µL of PB, 2 µL of 46 mg
-
1
-1
mL BOD in PB, and 2 µL of 7 mg mL PEGDGE in water. A 5-µL
aliquot of the mixed solution was pipetted onto the mounted hydrophilic
carbon cloth, which was promptly wetted and penetrated by the solution.
The electrodes were cured for at least 18 h at room temperature before
they were used.
Instrumentation and Cell. The measurements were performed using
a bipotentiostat (CH-Instruments, electrochemical detector model
CHI832) and a dedicated computer. The temperature was controlled
with an isothermal circulator (Fisher Scientific, Pittsburgh, PA). The
2 2
dissolved O concentration was monitored with an O electrode
purchased from BAS (West Lafayette, IN). The electrodes were rotated
using a Pine Instruments rotator (Austin, TX). The measurements were
carried out in a water-jacketed electrochemical cell at 37.5 °C containing
5
0 mL of PBS. At the start of the experiments, argon was bubbled
through the solution for a least 15 min, followed by oxygen. To maintain
a fixed volume of solution in the cell, the bubbled gases were
presaturated with water by passage through a bubbler, which also
contained PBS. The potentials were measured versus a commercial Ag/
AgCl (3 M KCl) reference electrode. The counter electrode was a
platinum wire (BAS). In the coulometric measurements, the scan rate
Figure 2. Cyclic voltammogram of the “wired” bilirubin oxidase-coated
carbon cloth electrode under argon. PBS (0.15 M NaCl, pH 7.4; 20 mM
-1
phosphate) buffer, 37.5 °C, and 50 mV s scan rate. Total loading, of the
44.6 wt % BOD, 48.5 wt % redox polymer, 6.9 wt % PEGDGE cross-
-
2
2
linker composite, 0.6 mg cm . Current densities based on the 0.126-cm
geometrical area of the 4-mm-diameter carbon cloth disk.
-
1
was 1 mV s
.
BOD Assay. The absorption spectra of the BOD solutions were
measured at 25 °C with an Agilent 8453 UV-visible spectrophotometer
fraction in the catalyst, ∆Ep and Ewhm did not vary with the
PEGDGE weight fraction in the “wired” BOD electrodes. As
will be shown below, the kinetic limit of their current density
depended, however, on the weight fraction of the PEGDGE
4
1
following the procedure of Hirose. The concentration of BOD was
-
1
calculated using its reported molar absorption coefficient of 3870 M
cm at 610 nm.
-
1
41,42
-
1
cross-linker. Upon 4-h cycling of the potential at 50 mV s
Results
between 0.0 and +0.5 V versus Ag/AgCl, the heights of the
voltammetric peaks of the electrodes rotating at 1000 rpm
decreased only by less than 5%, at 37.5 °C.
The open circuit potential of a vitreous carbon electrode on
which BOD (without redox polymer) was adsorbed was +360
mV versus Ag/AgCl in pH 7.4 PBS under 1 atm O2 at 37.5 °C
or -196 mV versus the potential of the reversible O2/H2O under
the same conditions. The open circuit potential of a similar
electrode, but with a thin adsorbed film of the electrostatic
adduct of the redox polymer and BOD, was remarkably higher,
The reproducibility of the O2 electroreduction currents of the
electrodes rotating at 1000 rpm was (10% or better both for
different batches (n ) 4) and within batches (n ) 12). When
the electrodes were rotated at 4000 rpm, the reproducibility was
only (28% between the batches and (20% within the batches.
Because two-electron BOD-catalyzed electroreduction of O2
+530 mV versus Ag/AgCl, only -26 mV versus the potential
4
4,45
to H2O2
reduction, H2O2 production was tested for using RRDEs having
wired” BOD carbon cloth disks and platinum rings. If H2O2
might compete with the desired four-electron
of the reversible O2 /H2O electrode. No spectroscopic evidence
was found for complex formation between the copper ions of
BOD and the polymer: When the water-soluble copolymer of
acrylamide and N-vinylimidazole was added to the BOD
solution in PB, the spectrum of the enzyme did not change.
Figure 2 shows the cyclic voltammogram of the composite
“
were produced on the disk, it would have been detected at the
ring poised at +0.750 mV versus Ag/AgCl, the H2O2 being
electrooxidized to O2 on the platinum ring at this potential. (A
similar experiment was performed to investigate the formation
“
wired” BOD-coated carbon cloth electrode under argon at 50
-
1
of H2O2 during the electroreduction of O2 on disks of RRDE
mV s in PBS buffer at 37.5 °C. The coating of the electrode
electrodes modified with cytochrome c and analogs.4
6-49
).
consisted of 44.6 wt % BOD, 48.5 wt % polymer, and 6.9 wt
When O2 was electroreduced on the disk at increasing current
%
PEDGE. Its voltammogram is characteristic of a polymer-
-
2
densities between 0 and 5 mA cm , the ring current remained
unchanged and was negligibly small (Figure 3A), showing that
H2O2 was not produced on the disk. Furthermore, catalase,
bound osmium complex with an apparent redox potential of
-
1
+
350 mV versus Ag/AgCl. At 1 mV s scan rate, the
voltammogram exhibited a symmetrical wave, with only slight
separation of the oxidation and reduction peaks (∆EP ) 5 mV).
(
43) Bard, A. J.; Faulkner, L. R. Electrochemical Methods. Fundamentals and
Applications; John Wiley and Sons: New York, 2001.
-
1
At 50 mV s the separation was substantially increased (∆EP
53 mV). The width of the peaks at half-height, Ewhm, was 90
(
44) Shoham, B.; Migron, Y.; Riklin, A.; Willner, I.; Tartakovsky, B. Biosens.
Bioelectron. 1995, 10, 341-352.
)
mV, close to the theoretical width of 90.6 mV for an ideal
Nernstian one-electron-transfer reaction. In contrast with wired
laccases, where ∆Ep and Ewhm increased with the PEDGE weight
(45) Wang, J.; Ozsoz, M. Electroanalysis 1990, 2, 647-650.
(
46) Collman, J. P.; Fu, L.; Herrmann, P. C.; Zhang, X. Science 1997, 275,
949-951.
(47) Collman, J. P.; Fu, L.; Herrmann, P. C.; Wang, Z.; Rapta, M.; Broring,
M.; Schwenninger, R.; Boitrel, B. Angew. Chem., Int. Ed. Engl. 1998, 1998,
3397-3400.
(
41) Hirose, J.; Inoue, T.; Sakuragi, H.; Kikkawa, M.; Minakami, M.; Morikawa,
T.; Iwamoto, H.; Hiromi, K. Inorg. Chim. Acta 1998, 273, 204-212.
42) Hiromi, K.; Yamaguchi, S.; Sugiura, Y.; Iwamoto, H.; Hirose, J. Biosci.
Biochem. 1992, 56, 1349-1352.
(48) Collman, J. P.; Rapta, M.; Broring, M.; Raptova, L.; Schwenninger, R.;
Boitrel, B.; Fu, L.; L’Her, M. J. Am. Chem. Soc. 1999, 121, 1387-1388.
(49) Pardo-Yissar, V.; Katz, E.; Willner, I.; Kotlyar, A. B.; Sanders, C.; Lill,
H. Faraday Discuss. 2000, 116, 119-134.
(
6482 J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002

O
2
Cathode Operating in Physiological Solution
A R T I C L E S
Figure 4. Dependence of the current density on the BOD enzyme weight
-1
percentage. Electrode poised at +300 mV vs Ag/AgCl, 1 mV s scan rate,
1000 rpm, and 1 atm O2. Other conditions as in Figure 2.
Figure 3. (A) Polarization of the cathode under 1 atm O2, rotating at 1000
top curve), 2300 (middle curve), and 4000 rpm (bottom curve), in PBS
(
-
1
buffer at 37.5 °C. Scan rate, 1 mV s . Other conditions as in Figure 2. (B)
Current density versus the square root of the angular velocity (ω1 ) for the
electrode poised at +300 mV vs Ag/AgCl in PBS at 37.5 °C (open circles)
and the current density calculated by Levich equation (solid line) (see text).
Conditions, other than angular velocities and the 1 atm O2 pressure, as in
Figure 2.
/2
Figure 5. pH dependence of the steady-state current density under 1 atm
O2 for the electrode poised at +300 mV vs Ag/AgCl, 1000 rpm. Other
conditions as in Figure 2.
which would have decomposed any H2O2 that might have been
produced into O2, had no effect on the ring current, confirming
that H2O2 was not a byproduct of the electroreduction of O2 to
water. After 10 h of O2 electroreduction at ∼1 mA current, the
The precipitation is attributed to the formation of a charge-
neutral electrostatic adduct between the cationic polymer and
the anionic enzyme (pI ) 4.1). In experiments where the weight
percentage of the cross-linker was varied, while the BOD/redox
polymer weight ratio was fixed at 1:1, the optimal PEGDGE
weight percentage was found to be 6.9. The resulting optimal
catalyst was composed of 44.6 wt % BOD, 48.5 wt % polymer,
and 6.9 wt % PEDGE at 0.6 mg cm total loading. The effect
of the total loading on the kinetic limit of the catalytic current
was not optimized, because the optimum would have varied
with the ratios of the components.
5
0-mL solution of the cell tested negative for H2O2 in a
colorimetric ABTS-horseradish peroxidase test that would have
detected 1 µM H2O2.
Figure 3B shows the dependence of the current density on
-
2
the square root of the angular velocity of the rotating electrode,
1
/2
ω . Up to 3500 rpm, where the current density reached 8.8
-
2
1/2
mA cm , the current increased linearly with ω in accordance
with the Levich equation, showing that it was O2 transport-
limited under 1 atm O2 when the electrode was poised at +300
mV versus Ag/AgCl. Above 3500 rpm, the current continued
5
0-52
Unlike in the cases of wired glucose oxidase
and of wired
1
9
laccases, cross-linking by periodate oxidation of the BOD
oligosaccharides did not lead to useful wired BOD electrodes,
even though BOD contains 6.3 wt % carbohydrate.6
1/2
to increase with ω , but no longer linearly, until it reached
the kinetic limit of 9.1 mA cm .
-
2
The pH dependence of the steady-state current density of O2
electroreduction was measured with the electrode poised at +300
mV versus Ag/AgCl in 0.15 M NaCl while the electrode rotated
at 1000 rpm. Phosphate, borate, citrate, or Tris was added at
The optimal composition of the electrocatalyst was deter-
mined for electrodes rotating at 1000 rpm and poised at +300
mV versus Ag/AgCl. In the first group of experiments, the cross-
linker (PEDGE) was fixed at 6.9 wt %, and the total loading of
2
0 mM concentration to maintain the desired pH. The de-
-
2
all film components was fixed at 0.6 mg cm . Figure 4 shows
pendence of the current on the pH is shown in Figure 5. As
seen in the figure, the current density increased with pH until
the dependence of the current density of BOD through the 20-
7
0 wt % range. From 20 to 45 wt %, the current density
increased with the weight percentage of BOD, reaching 4.6 mA
cm at 45.2 wt %. Above 50 wt % BOD, the current density
declined rapidly. At 60 wt % BOD, precipitation was observed.
(50) Binyamin, G.; Heller, A. J. Electrochem. Soc. 1999, 146, 2965-2967.
(51) Vreeke, M.; Maidan, R.; Heller, A. Anal. Chem. 1992, 64, 3084-3090.
-
2
(
52) Binyamin, G.; Cole, J.; Heller, A. J. Electrochem. Soc. 2000, 147, 2780-
2783.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002 6483

A R T I C L E S
Mano et al.
Figure 7. Temperature dependence of the current density. 1 atm O2,
electrode poised at +300 mV vs Ag/AgCl, 1000 rpm. Other conditions as
in Figure 2.
Figure 6. Dependence of the current density on the concentrations of
different anions under 1 atm O2 for the electrode poised at +300 mV vs
Ag/AgCl, 1000 rpm. Other conditions as in Figure 2.
it reached a plateau at pH 7.5 and then declined above pH 10.5.
In the pH 6-10.5 range, the current density was nearly
independent of pH, varying by less than (10%. Up to pH 9,
there was no irreversible change in the current characteristics;
above pH 10.5, the drop in the current density was irreversible.
Because copper-binding anions, particularly halide anions,
inhibit the laccases and because blood contains 0.14 M chloride,
the effect of added anions was investigated. In these experi-
ments, the electrode rotated at 1000 rpm and was poised at +
3
00 mV versus Ag/AgCl in a pH 7.4, 20 mM phosphate buffer
solution. The results are summarized in Figure 6. The current
density was nearly independent of the concentration of chloride,
added as NaCl, through the 0-1 M range. Above 1.5 M NaCl
concentrations, a decline in the current density was, nevertheless,
observed. When bromide was added, the current density declined
monotonically as the KBr concentration was raised from 0 and
Figure 8. Stability of the ‘wired’ enzyme carbon cloth electrode poised at
+300 mV vs Ag/AgCl under 1 atm O2, 300 rpm. Other conditions as in
Figure 2.
1
M. At 1 M KBr, the current density was about 25% lower
than in the absence of bromide. A steeper decline was again
observed at >1.5 M KBr concentration. Adding of fluoride (as
NaF) inhibited the electroreduction of O2, about 30% of the
current density being lost when the concentration of fluoride
was raised from 0 to 0.5 M and about 90% being lost when the
concentration was further raised to 1 M. Thiocyanate, added as
KCNS, strongly inhibited the electroreduction, the current
density dropping to 0 already at 0.25 M KCNS. The maximum
current density was about the same when the solutions were
buffered at pH 7 with Tris, acetate, or phosphate.
at 300 rpm, where the current was O2 transport limited. At 300
rpm, the shear stress acting on the rim of the 4-mm-diameter
electrode was 1.4 × 10 N m . The time dependence of the
current density of an electrode poised at + 300 mV versus Ag/
AgCl rotating at 300 rpm in PBS under 1 atm O2 at 37.5 °C is
shown in Figure 8. In the first 6 days of operation, the current
-
2
-2
-
2
dropped by less than 10% per day: The initial 2.4 mA cm
-
2
current density dropped to 1.3 mA cm after 6 days of
continuous operation. After storage of the dry electrodes for 1
month at 4 °C under air, 95 ( 3% of the initial current density
was retained. The presence of electrooxidizable blood constitu-
ents did not harm the electrodes. A transient 5% increase in the
current density was observed when either 0.1 mM ascorbate or
Figure 7 shows the temperature dependence of the current
density of the electrode poised at +300 mV versus Ag/AgCl,
rotating at 1000 rpm in PBS under 1 atm O2. The current density
increased with temperature up to 60 °C and then declined
rapidly. The increase was reversible only up to 50 °C, the
enzyme being denatured at higher temperatures. The Arrhenius
plot (not shown) of the temperature dependence yielded an
0
.1 mM ascorbate and 0.17 mM acetaminophen was added, and
a transient 25% increase in current density was observed with
.48 mM urate in the solution.
0
-
1
activation energy of Eact ) 28.2 kJ mol . The observed
Discussion
activation energy for the denaturing of the enzyme was 77 kJ
-
1
mol .
When rotating at 4000 rpm, where the shear stress on the
Potentials. Xu et al. measured spectrophotometrically the
redox potential of an aerated pH 5.3 solution of recombinant
M. Verrucaria BOD at 20 °C, reporting +260 mV versus Ag/
AgCl or -440 mV relative to the potential of the reversible
-
2
rim on the fiber cloth was 0.7 N m , the electrodes were
mechanically unstable. For this reason, their stability was tested
6484 J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002

O
2
Cathode Operating in Physiological Solution
A R T I C L E S
O2/H2O electrode at the same pH.53 In pH 7.8 50 mM Tris
through the redox polymer. The ratio also depends on the nature
of the bond between the wire and the enzyme, which determines
the maximum electron current from the redox polymer to BOD.
In the absence of bonding, the BOD and the redox polymer
would phase-separate, the entropy of mixing of two macromol-
ecules being too low prevent their separation. As discussed
above, intimate contact between the electrostatically bound BOD
and its wire lowers the kinetic resistance for electron transfer
from the wire to BOD.
buffer, Shimizu et al. reported a potential of +373 mV versus
3
4
Ag/AgCl at ambient temperature, -174 mV versus the
reversible O2/H2O electrode potential at this pH. The open circuit
potential of the O2 electrode made by adsorbing BOD on
vitreous carbon was +360 mV versus Ag/AgCl under 1 atm
O2 at 37.5 °C and at pH 7.4, consistent with the latter value.
This potential was -196 mV versus that of the reversible O2/
H2O electrode under these conditions. The open circuit potential
of the “wired” BOD electrode, made by cross-linking the
electrostatic adduct of BOD, a polyanion above pH 4.1, and
the redox polycation of Figure 1 on carbon cloth, was +530
mV versus Ag/AgCl, only 26 mV below that of the reversible
O2/H2O electrode, at pH 7.4 under 1 atm O2 and at 37.5 °C.
The cause of the remarkable increase in potential upon wiring
the BOD is not known. The redox potential of the PAA-PVI-
Cross-linking reduces the segmental mobility on which the
electron conduction in the redox polymer films depends.
Because electrons diffuse as they are transferred when reduced
and oxidized redox centers collide, the highest electron diffu-
sivities are being reached when the film is not cross-linked.56
However, in absence of cross-linking the films dissolve, and
when inadequately cross-linked, they are sheared off the rotating
+/2+
50
[Os(4,4′-dichloro-2,2′-bipyridine)2Cl]
“wire” (Figure 1) was
electrodes. With 6.9 wt % PEDGE cross-linker, the films were
+
350 mV versus Ag/AgCl or -180 mV versus the pH 7.4 open
mechanically stable at 300 rpm where the maximum shear stress
-
2
-2
circuit potential of the “wired” BOD electrode under 1 atm O2.
at the rims of the electrodes was 1.4 × 10 N m . At this
-
2
Voltammetric Characteristics. The slight separation of the
angular velocity, the current density was 2.4 mA cm
.
-
1
voltammetric peak heights at 1 mV s scan rate is indicative
of a reversible surface-bound couple. The linear dependence of
the peak heights on the scan rate also shows that the redox
couples of the polymer are surface-confined.43 The voltammetric
peaks were narrower and less separated than those in the earlier
PVI-Os(tpy)(dme-bpy) -based laccase electrode, where ∆Ep
was 120 mV.17 The difference is attributed to faster charge
transport through the “wire” of this study. Repetitive cycling
over a 4-h period at 37.5 °C did not change the shape of the
voltammograms of the electrodes rotating at 1000 rpm, and their
peak currents decreased only by less than 5%.
Reproducibility. When the electrodes rotated at 1000 rpm,
the apparent batch-to-batch reproducibility, as well as the within-
the batch reproducibility, of the O2 electroreduction currents
was better than (10%. This reproducibility does not necessarily
imply that the composite catalysts were identical within (10%,
because at 1000 rpm the current was limited by O2 transport,
not by the kinetics of the electrocatalyst. When the electrodes
were rotated at 4000 rpm, where the currents were already
limited by the kinetics of the catalytic films, but the electrodes
were already mechanically unstable, the reproducibility was only
Polarization. Electroreduction of O2 started at the open circuit
potential of the wired BOD electrode, +530 mV versus Ag/
AgCl. When the electrode rotated at 4000 rpm where, as will
be shown below, the current was no longer O2 transport limited,
-
2
the current density was 6.8 mA cm at +380 mV versus Ag/
AgCl (-176 mV versus the potential of the reversible O2/H2O
electrode at the same pH) and the current density reached 9.1
2+/3
-
2
mA cm at +300 mV versus Ag/AgCl, -276 mV versus the
-
2
reversible potential. The value of 6.8 mA cm at -176 mV
versus the potential of the reversible O2/H2O electrode represents
a greater than 13-fold increase over current density reported by
29
Tsujimara et al., whose solution was not a physiological buffer,
and whose cathodes operated only for a few hours.
Mass Transport and Kinetics. The Levich-Koutecky plot57
for electrode rotating at an angular velocity ω (Figure 3B) shows
2
/3 -1/6 1/2 1/2
that the Levich equation I ) 0.62nFAcD ν ω ω is
rigorously obeyed up to 3500 rpm, where the current density is
-
2
8.8 mA cm . The correspondence between the measured
(circles) and the predicted currents (solid line) is exact for the
-
5
2
-1
O2 diffusivity D ) 2.4 × 10 cm s , the O2 solubility c )
-
6
-3
1.07 × 10 mol cm , and the kinematic viscosity ν ) 0.091
2
-1
(
28% between the batches and ( 20% within the batches.
cm s of the 0.15 M NaCl solution at 37.5 °C, when n, the
Composition. The optimal enzyme-to-polymer weight ratio
number of electrons electroreducing the O2, equals 4 and the
2
58-60
was near 1:1 (Figure 4). At a lower weight fraction of enzyme,
the current was limited by the BOD-catalyzed rate of O2
reduction. When the weight fraction of enzyme was higher, the
current was limited by the electronic resistance of the film, the
redox polymer being an electron conductor and the enzyme an
area A is 0.126 cm .
Above 3500 rpm, the current is
controlled by the kinetics of the electrocatalytic reaction. When
the electrode rotates at 4000 rpm and the current is kinetically
controlled, the measured Tafel slope is -122 mV/decade, the
theoretical slope for O2 electroreduction at 37.5 °C.61
Absence of Inhibition in Physiological Solution. The pH
dependence of the current density of the “wired” BOD electrode
differed from the pH dependence of the maximal turnover rate
Vmax of the dissolved enzyme. When the enzyme is dissolved,
5
4,55
insulator.
Above 60 wt % BOD, where the electrostatic
adduct precipitated, the increased resistance resulted in increased
separation of the voltammetric peaks. At the optimal weight
ratio, the upper limits of three currents, of the flow of electrons
from the electrode to the redox polymer, from the redox polymer
to BOD, and from BOD to O2 are equal. The optimal ratio
decreases when the maximum turnover rate of the enzyme is
higher and decreases when electrons diffuse more rapidly
(
(
(
56) Aoki, A.; Rajagopalan, R.; Heller, A. J. Phys. Chem. 1995, 99, 5102-
5105.
57) Bard, A. J.; Faulkner, L. R. Electrochemical methods: Fundamentals and
applications, 2nd ed.; John Wiley and Sons: New York, 2001; p 341.
58) Battino, R. Oxygen and Ozone, Solubility data series (Pergamon, Oxford,
1981), vol. 7.
(
(
(
53) Xu, F.; Shin, W.; Brown, S. H.; Walhleithner, J. A.; Sundaram, U. M.;
(59) Lobo, V. M. M. Handbook of electrolyte solutions (Elsevier, Amsterdam,
1989).
Solomon, E. I. Biochim. Biophys. Acta 1996, 1292, 303-311.
54) Ohara, T. J.; Rajagopolan, R.; Heller, A. Anal. Chem. 1993, 65, 3512-
(60) Weast, R. C. Handbook of Chemistry and Physics (CRC Press, Boca Raton,
1986).
3
517.
55) Ohara, T. J.; Rajagopolan, R.; Heller, A. Anal. Chem. 1994, 66, 2451-
(61) Bard, A. J.; Faulkner, L. R. Electrochemical methods: Fundamentals and
applications, 2nd ed.; John Wiley and Sons: New York, 2001; p 103.
2
457.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002 6485

A R T I C L E S
Mano et al.
Vmax was high only through the pH 6-8 range. Vmax declined
rapidly at both higher and lower pH, only one-third of Vmax
being retained at pH 9.26 In contrast, the “wired” BOD electrode
retained its high current density through the pH 6-10 range.
Importantly, the variation of the current was negligibly small
within the pH 7.35-7.45 range of normal human arterial blood
with the electrode rotating at 1000 rpm under 1 atm O2, where
the current was O2 transport controlled. Because the current was
O2 transport limited at 1000 rpm, the 28.2 kJ mol value
-1
represents the activation energy for diffusion of oxygen in the
6
7
-1
solution. The 77 kJ mol activation energy for current loss
is similar to that for the thermal denaturation of other enzymes.68
Stability. At high angular velocity (>1000 rpm), the electrode
was mechanically unstable, its catalytic film being sheared off.
When the electrode, poised at +300 mV versus Ag/AgCl, was
rotated at 300 rpm under 1 atm O2 at 37.5 °C, it lost less than
10% of its current per day of continuous operation for 6 days
(Figure 8). The “wired” BOD electrode was far more stable
than suggested by Tanaka’s study of the stability of the dissolved
enzyme, showing loss of half the activity after 1 h in phosphate
(Figure 5). Above pH 10.5, the electrode was irreversibly
damaged because of enzyme denaturation. The observed damage
above pH 10.5 is consistent with results of Samejima et al. who
found, by tracking the circular dichroism as a function of pH,
that the enzyme is denatured upon ionization of its tyrosine
6
2
residues. The voltammetric characteristics of the redox
polymer did not change up to pH 11.
Laccases are strongly inhibited by halide anions,63-65 includ-
ing the chloride anion, because these bind with their catalytic
type II Cu centers and prevent electron transfer from their type
I Cu to their type III Cu clusters.2 The current density of the
2
6
buffer at 37 °C. It was also more stable than that of an
immobilized BOD electrode used in the clinical monitoring of
the concentration of bilirubin. The half-life of the clinical
bilirubin-monitoring electrode was much shorter than that of
6,41
“
wired” laccase electrodes declined at pH 5 by 60% when the
19
69
chloride concentration was raised from 0 to 0.1 M. In contrast,
the current density of the “wired” BOD electrode declined only
by 6%. The loss remained small even in 1 M NaCl at pH 7.4
the cathode of this study, only 17 h at 37 °C and 8 h at 40
4
4
°C.
The most readily cathode electooxidized constituents of blood,
urate, ascorbate and acetaminophen, at their physiological
(
Figure 6). Other copper-binding anions inhibited the electro-
-
70
reduction of O2, the inhibition declining in the order CNS .
F
agreement with the order CNS > F . Cl > Br of solution-
phase inhibition of the free enzyme reported by Hirose et al.
for BOD and by Xu et al.53,63 for laccases. Significantly, the
concentrations, did not damage the cathode poised at +300
-
-
-
. Br >Cl (Figure 6). This order is only in partial
mV/Ag/AgCl in PBS buffer at 37.5 °C. When 0.1 mM ascorbate,
or 0.1 mM ascorbate and 0.17 mM acetaminophen was added,
the current transiently decreased by 5%, and when 0.48 mM
urate was added, the decrease was 25%. While Binyamin et
-
-
-
-
4
1
3
5
inhibition of the “wired” BOD cathode by chloride at its 0.096-
al. showed that the electrooxidative polymerization of urate
in a redox polymer film with a higher density of cationic sites
and thus a higher urate concentration in the film caused
degradation of its electrocatalytic properties, no such damage
was observed in the present films.
0
.106 M concentration in normal human blood was so small
that physiological variation in chloride concentration will not
affect the O2 electroreduction current. The electrode differed
in this respect from the dissolved BOD-based cathode of
2
9
Tsujimura et al., the current of which depended strongly on
the concentration of KCl in the 0-0.1 M range. The observed
partial loss of O2 electroreduction current at very high chloride
concentration (>1.5 M NaCl) is not attributed to specific binding
of chloride to copper centers of BOD. Its likely cause is the
screening of the charges of the electrostatically bound poly-
cationic and polyanionic segments of the cross-linked redox
polymer-BOD adduct at high ionic strength, causing their
dissociation and making the “wiring” of the enzyme ineffec-
Conclusion
The study describes the first electrode on which O2 is
electroreduced to water physiological conditions at a current
-
2
density of 9.1 mA cm at a potential -0.26 V relative to the
reversible potential of the O2/H2O electrode. The electrode has
no leachable components, making it suitable for use in flow
systems, and with a yet to be added thin bioinert film, for use
in animals. It could well serve as the cathode of a miniature
biofuel cell that might power implanted sensors and actuators
for about 1 week.
5
4,55,66
tive.
Temperature Dependence. The O2 electroreduction current
increased with the temperature up to 60 °C when the rate of
Acknowledgment. The study was supported by the Office
of Naval Research (Grant N00014-02-1-0144). The authors
thank Ting Chen, Keith A. Friedman, Scott Calabrese Barton,
Gary Binyamin, and Vonda Totten for their advice and
assistance.
-
1
increase was about 10 °C h and declined rapidly when the
temperature exceeded 60 °C. The decline is attributed to the
denaturation of the enzyme (Figure 7). The apparent stability
of the “wired” enzyme electrode was better than that of the
dissolved enzyme, for which an optimal temperature of 40 °C
was reported.2
5,26
JA025874V
The apparent activation energy for O2 elec-
-
1
troreduction was Eact ) 28.2 kJ mol for the pH 7.4 solution
(
67) Perry, R. H.; Green, D. W. Chemical Engineers Handbook; McGraw-Hill
Book Co.: New York, 1984.
(
62) Samejima, T.; Wu, C. S.; Shiboya, K.; Kaji, H.; Koikeda, S.; Ando, K.;
(68) Rob, A.; Hernandez, M.; Ball, A. S.; Tuncer, M.; Arias, M. E.; Wilson,
M. T. Appl. Biochem. Biotechnol. 1997, 62, 159-174.
(69) Sung, C.; Lavin, A.; Klibanov, A. M.; Langer, R. Biotechnol. Bioenerg.
1986, 28, 1531-1539.
Yang, J. T. J. Protein Chem. 1994, 13, 307-313.
(
(
(
(
63) Xu, F. Biochemistry 1996, 35, 7608-7614.
64) Xu, F. J. Biol. Chem. 1997, 272, 924-928.
65) Xu, F. Appl. Biochem. Biotechnol. 2001, 95, 125-133.
66) Heller, A. J. Phys. Chem. 1992, 96, 3579-3587.
(70) Csoregi, E.; Schmidtke, D. W.; Heller, A. Anal. Chem. 1995, 34, 1240-
1244.
6486 J. AM. CHEM. SOC.
9
VOL. 124, NO. 22, 2002