Esther Edwardes Moore1, Virgil Andrei1, Sónia Zacarias2, Inês A C Pereira2, Erwin Reisner1. 1. Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, United Kingdom. 2. Instituto de Tecnologia Química e Biológica António Xavier, Universidade Nova de Lisboa, Av. da Republica, 2780-157 Oeiras, Portugal.
Abstract
Lead halide perovskite solar cells are notoriously moisture-sensitive, but recent encapsulation strategies have demonstrated their potential application as photoelectrodes in aqueous solution. However, perovskite photoelectrodes rely on precious metal co-catalysts, and their combination with biological materials remains elusive in integrated devices. Here, we interface [NiFeSe] hydrogenase from Desulfovibrio vulgaris Hildenborough, a highly active enzyme for H2 generation, with a triple cation mixed halide perovskite. The perovskite-hydrogenase photoelectrode produces a photocurrent of -5 mA cm-2 at 0 V vs RHE during AM1.5G irradiation, is stable for 12 h and the hydrogenase exhibits a turnover number of 1.9 × 106. The positive onset potential of +0.8 V vs RHE allows its combination with a BiVO4 water oxidation photoanode to give a self-sustaining, bias-free photoelectrochemical tandem system for overall water splitting (solar-to-hydrogen efficiency of 1.1%). This work demonstrates the compatibility of immersed perovskite elements with biological catalysts to produce hybrid photoelectrodes with benchmark performance, which establishes their utility in semiartificial photosynthesis.
Lead halide perovskite solar cells are notoriously moisture-sensitive, but recent encapsulation strategies have demonstrated their potential application as photoelectrodes in aqueous solution. However, perovskite photoelectrodes rely on precious metalco-catalysts, and their combination with biological materials remains elusive in integrated devices. Here, we interface [NiFeSe] hydrogenase from Desulfovibrio vulgaris Hildenborough, a highly active enzyme for H2 generation, with a triple cation mixed halide perovskite. The perovskite-hydrogenase photoelectrode produces a photocurrent of -5 mA cm-2 at 0 V vs RHE during AM1.5G irradiation, is stable for 12 h and the hydrogenase exhibits a turnover number of 1.9 × 106. The positive onset potential of +0.8 V vs RHE allows its combination with a BiVO4 water oxidation photoanode to give a self-sustaining, bias-free photoelectrochemical tandem system for overall water splitting (solar-to-hydrogen efficiency of 1.1%). This work demonstrates the compatibility of immersed perovskite elements with biological catalysts to produce hybrid photoelectrodes with benchmark performance, which establishes their utility in semiartificial photosynthesis.
As a globally abundant and economical
energy source, solar energy is the fastest growing renewable alternative
to fossil fuels.[1,2] Artificial photosynthesis uses
sunlight for the production of renewable chemical fuels, so-called
solar fuels, thus addressing the intermittency limitations of photovoltaic
(PV) technologies.[3,4] Solar fuel synthesis can be achieved
by direct coupling of an efficient light absorber to a fuel-producing
catalyst.[5,6] Organic–inorganic lead halide perovskites
have received much attention due to their low production costs and
promising PV cell efficiencies, currently reaching up to 25.2%.[2,7−10] However, moisture, air, and temperature instability has challenged
the use of perovskites in photoelectrochemical (PEC) devices.[11,12] Encapsulation layers such as eutectic metal alloys, metal foils,
and epoxy resin have improved the operation lifetime of solution-immersed
perovskite-based photoelectrodes from seconds to hours.[11,13−17] However, all H2-evolving PEC perovskite photocathodes
have so far employed high-cost, low-abundance Pt nanoparticles as
the co-catalyst.Semiartificial photosynthesis combines the
evolutionarily optimized
activity of biological catalysts, such as isolated enzymes, with synthetic
photoabsorbers.[18−21] Hydrogenases (H2ases) are reversible and highly efficient
H2 production enzymes with a per-active-site activity matching
that of Pt.[22−24] The integration of H2ase with Si and Cu2O photocathodes has previously been achieved,[25−29] but the combination with an immersed lead halide perovskite has
remained inaccessible due to the moisture sensitivity of this photoabsorber
and difficulty of achieving a productive enzyme–photoabsorber
interface.Here, a perovskite–H2ase photocathode
is presented,
realized by an encapsulation system that protects the photoabsorber
and provides a biocompatible, bespoke porous TiO2 scaffold
for the enzyme. This semiartificial photocathode enabled combination
with a BiVO4 water oxidation photoanode for bias-free,
tandem PEC water splitting into H2 and O2 (Figure ).
Figure 1
Schematic representation
of the tandem PEC cell consisting of a
FM-encapsulated perovskite photocathode with H2ase integrated
into an IO-TiO2 layer and a BiVO4 photoanode.
TiCo refers to the water oxidation layer precursor: [Ti4O(OEt)15(CoCl)]. PCBM: [6,6]-phenyl-C61-butyric
acid methyl ester. PEIE: polyethylenimine.
Schematic representation
of the tandem PEC cell consisting of a
FM-encapsulated perovskite photocathode with H2ase integrated
into an IO-TiO2 layer and a BiVO4 photoanode.
TiCo refers to the water oxidation layer precursor: [Ti4O(OEt)15(CoCl)]. PCBM: [6,6]-phenyl-C61-butyric
acid methyl ester. PEIE: polyethylenimine.Optimized cesium formamidinium methylammonium (CsFAMA)
triple cation
mixed halide perovskite devices with a Field’s metal (FM) protection
layer were assembled and characterized as previously reported (Figure ; see SI Experimental Procedures and Figure S1 for details).[15] Enzymes
have been integrated with high loading into hierarchically structured,
macro- and mesoporous, inverse opal (IO) metaloxide scaffolds.[25,30,31] TiO2 was selected
in this study for its stability and conductivity under reducing conditions
as well as its ability to form a biocompatible interface with enzymes.[25,32,33] The high-temperature (>100
°C)
sensitivity of the perovskite prevented in situ annealing of the IO-TiO2 directly on the FM surface. Therefore, anatase TiO2 nanoparticles (∼21 nm Ø) were first co-assembled with
polystyrene beads (750 nm Ø) on Ti foil and annealed at 500 °C
to give Ti|IO-TiO2 (Figure S2). The geometrical surface area of the IO-TiO2 scaffold
was 0.28 cm2 with an IO-TiO2 film thickness
of 15 μm. The Ti|IO-TiO2 was then joined to the protected
perovskite by briefly melting the FM sheet with a Peltier thermoelectric
element (at ∼70 °C), and an epoxy resin was used to seal
the edges to give the encapsulated PV-integrated photocathode: PVK|IO-TiO2 [FTO-glass|NiO|perovskite|PCBM|PEIE|Ag|FM|Ti|IO-TiO2] (Figure ).A [NiFeSe] H2ase from Desulfovibrio vulgaris Hildenborough (DvH) was selected for its considerable
H2 evolution activity compared to that of DvH [NiFe] H2ase, and was purified and characterized as
previously reported.[23,33−37] The selenocysteine residue (Sec489) in the active
site (Figure S3) causes improved O2 tolerance,[35,37−40] which is beneficial for its application
in overall water splitting. The [NiFeSe] H2ase (5 μL,
50 pmol) was drop-cast onto Ti|IO-TiO2 and left to saturate
the film for 30 min in a N2 atmosphere. Protein film voltammetry
of the Ti|IO-TiO2|H2ase electrode in a three-electrode
configuration demonstrated that proton reduction occurred with minimal
overpotential, indicative of efficient charge transfer at the TiO2–H2ase interface (Figure S4). The quality of the interface can be attributed to the
well-known strength of protein binding to TiO2, an effect
that may be further accentuated by polarization of the TiO2 surface.[25,33,41] The Ti|IO-TiO2|H2ase electrode displayed current
densities of −2.5 mA cm–2 with high stability
for several hours at an applied potential (Eapp) of −0.5 V vs RHE under N2, including
some robustness in the presence of O2. A Faradaic efficiency for H2, FEH, of 78% after 24 h was determined
by gas chromatography. The Eapp of −0.5
V vs RHE was applied to reflect the estimated perovskite photovoltage
of 0.9 V in the PEC experiments, where +0.4 V vs RHE has been applied
(see below).Protein–film photoelectrochemistry of the
PVK|IO-TiO2|H2ase photocathode (three-electrode
configuration,
H2ase integrated as above) was conducted at 25 °C
under chopped simulated solar light irradiation (100 mW cm–2, AM1.5G). The photocathode was irradiated from the back, which prevented
photoexcitation of TiO2. Linear sweep voltammetry (LSV)
of the assembled PVK|IO-TiO2|H2ase electrode
showed a cathodic onset potential at +0.8 V vs RHE and a photocurrent
density of approximately −5 mA cm–2 at 0
V vs RHE (Figure a).
Figure 2
Photoelectrochemistry
of a biohybrid photocathode. (a) Representative
LSV of PVK|IO-TiO2|H2ase (blue), PVK|IO-TiO2 (green), and PVK (light blue) electrodes with chopped illumination
at a scan rate of 10 mV s–1. Arrow indicates start
of scan. (b) Representative CPPE at Eapp = +0.4 V vs RHE, with a dark period lasting 5 min following every
10 min of light exposure. (c) Mean (N = 3) H2 evolution from CPPE quantified by gas chromatography. Conditions:
MES (50 mM, pH 6.0), KCl (50 mM), DvH [NiFeSe] H2ase (50 pmol), simulated solar light back-irradiation (AM1.5G,
100 mW cm–2), N2 atmosphere, 25 °C.
Photoelectrochemistry
of a biohybrid photocathode. (a) Representative
LSV of PVK|IO-TiO2|H2ase (blue), PVK|IO-TiO2 (green), and PVK (light blue) electrodes with chopped illumination
at a scan rate of 10 mV s–1. Arrow indicates start
of scan. (b) Representative CPPE at Eapp = +0.4 V vs RHE, with a dark period lasting 5 min following every
10 min of light exposure. (c) Mean (N = 3) H2 evolution from CPPE quantified by gas chromatography. Conditions:
MES (50 mM, pH 6.0), KCl (50 mM), DvH [NiFeSe] H2ase (50 pmol), simulated solar light back-irradiation (AM1.5G,
100 mW cm–2), N2 atmosphere, 25 °C.Controlled potential photoelectrolysis (CPPE) was
conducted at
+0.4 V vs RHE, and gas chromatography was used to quantify H2 evolution yields. CPPE demonstrated the stability of the photocathode,
which consistently achieved 12 h of catalysis (Figure b). Failure of the enzyme–photocathode
after 12 h was likely due to water influx into the encapsulated perovskite,
consistent with previous reports (Figure S5).[13,15] The stability of the equivalent PVK-Pt device
was found to be comparable, supporting failure of the perovskite as
the limit to longevity (Figure S6). The
H2ase electrode generated 258 ± 55 μmolH cm–2 of H2, whereas
the enzyme-free electrode produced <1 μmolH cm–2 (Figure c). The FEH of PVK|IO-TiO2|H2ase after 14 h was (91 ± 1.5)% with a H2ase-based turnover number (TONH) of
1.9 × 106 and turnover frequency (TOFH) of 95 s–1.Bias-free tandem water
splitting has long been a desirable goal
for PEC cells.[25,31,42,43] Here a BiVO4-based water oxidation
photoanode was prepared by electrodeposition of BiOI, then drop-casting
and annealing a vanadium precursor, and finally spin-coating a layer
of a cobalt-containing co-catalyst, as previously reported.[15,44] PEC analysis of the photoanode (three-electrode setup; Figure S7) gave an onset potential of +0.1 V
vs RHE and a current density of 2.4 mA cm–2 at +1.23
V vs RHE.The positive onset potential of the PVK|IO-TiO2|H2ase photocathode is essential for combination
with the BiVO4 photoanode to assemble a tandem water splitting
PEC device.
The BiVO4 photoanode has been shown to absorb wavelengths
below 500 nm and therefore limits the perovskite to absorption at
500–800 nm.[15] Nevertheless, the
BiVO4 photoanode remains the current-limiting absorber
(Figure a). The robustness
of the [NiFeSe] H2ase towards O2 (Figure S4) provided the possibility to assemble
a “semiartificial leaf”, where the photoelectrodes were
not separated into two compartments by a membrane. The BiVO4||PVK|TiO2|H2ase tandem cell (Figure ) was prepared and PEC analysis
undertaken in a single-compartment with illumination through the front
of the BiVO4 photoanode.
Figure 3
Photoelectrochemistry of the tandem device.
(a) Representative
LSV of PVK|TiO2|H2ase (blue) and BiVO4 (green) electrodes with chopped illumination, forward scan, 10 mV
s–1 scan rate, showing the absolute current densities.
(b) Representative stepped potential chronoamperometry of BiVO4||PVK|TiO2|H2ase (blue) and H2ase-free BiVO4||PVK|TiO2 (green) tandem cells
from Uapp = −0.6 to +0.3 V. The
current density at Uapp = 0.0 V has been
highlighted. (c) Representative CPPE of BiVO4||PVK|TiO2|H2ase (blue) and H2ase-free BiVO4||PVK|TiO2 (green) tandem cells at Uapp = 0.0 V, with a dark period lasting 5 min following
every 10 min of light exposure. (d) Mean (N = 3)
H2 (dotted line with measurement points) and O2 (solid line) evolution from CPPE repeats. Conditions: MES (50 mM,
pH 6.0), KCl (50 mM), DvH [NiFeSe] H2ase
(50 pmol), simulated solar light irradiation (AM1.5G, 100 mW cm–2), N2 atmosphere, 25 °C.
Photoelectrochemistry of the tandem device.
(a) Representative
LSV of PVK|TiO2|H2ase (blue) and BiVO4 (green) electrodes with chopped illumination, forward scan, 10 mV
s–1 scan rate, showing the absolute current densities.
(b) Representative stepped potential chronoamperometry of BiVO4||PVK|TiO2|H2ase (blue) and H2ase-free BiVO4||PVK|TiO2 (green) tandem cells
from Uapp = −0.6 to +0.3 V. The
current density at Uapp = 0.0 V has been
highlighted. (c) Representative CPPE of BiVO4||PVK|TiO2|H2ase (blue) and H2ase-free BiVO4||PVK|TiO2 (green) tandem cells at Uapp = 0.0 V, with a dark period lasting 5 min following
every 10 min of light exposure. (d) Mean (N = 3)
H2 (dotted line with measurement points) and O2 (solid line) evolution from CPPE repeats. Conditions: MES (50 mM,
pH 6.0), KCl (50 mM), DvH [NiFeSe] H2ase
(50 pmol), simulated solar light irradiation (AM1.5G, 100 mW cm–2), N2 atmosphere, 25 °C.The two-electrode device achieved a current density
of 1.1 mA cm–2 under bias-free conditions (Uapp = 0.0 V), and stepped potential chronoamperometry
revealed
an onset potential of −0.6 V (Figure b). Bias-free CPPE showed a gradual decrease
in photocurrent over 8 h, which was attributed to slowly progressing
film loss due to enzyme inactivation, reorientation, or desorption
(Figure c). In agreement,
the current density returned to almost the initial value when a sacrificial
electron acceptor (methyl viologen) was added to the tandem PEC cell
after prolonged irradiation (Figure S8).
The peak FE of the device was (82 ± 3)% for H2 and
(50 ± 8)% for O2 (Figure d, FE over time; Figure S9). The lower FE for O2 can be attributed to some
O2 reduction at the photocathode leading to lower amounts
of O2 detected. The solar-to-hydrogen efficiency (STH)
was 1.1% (eq S1).The BiVO4||PVK|TiO2|H2ase cell
produced 21.2 ± 3.2 μmolH cm–2 and 9.0 ± 2.7 μmolO cm–2 after 8 h of CPPE, giving a H2:O2 ratio of 2.3. The PVK|IO-TiO2|H2ase photocathode (Figure S10) and BiVO4||PVK|TiO2|H2ase tandem device (Table , Figure S11) compare favorably with state-of-the-art H2 production PEC systems employing earth-abundant molecular
catalysts (synthetic and biological) in pH-benign aqueous solution
(see Tables S1 and S2 for details). Semiartificial
H2 evolution photocathodes have been previously reported
(Figure S10, color): a [NiFeSe] H2ase from Desulfomicrobium baculatum was introduced
onto a p-silicon (p-Si) photoabsorber via an IO-TiO2 scaffold,[25] whereas [FeFe] H2ases have been combined
with both p-type CuO2 and black-Si photoabsorbers.[26,27] Of the systems that employed small-molecule catalysts (Figure S10, gray scale), a Ni Dubois-type catalyst
applied to a p-Si photoabsorber and Fe-porphyrin and polymeric Co-based
catalysts combined with a GaP photocathode provide state-of-the-art
performances.[28,45,46] Previously reported tandem earth-abundant molecular catalyzed PEC
water splitting devices have utilized dye-sensitized p-type semiconductors
with cobaloxime H2 catalysts, resulting in STH values below
0.05% (Table ).[42] A semiartificial
tandem cell with a H2ase cathode was wired to an organic
dye–photosystem II photoanode, with a STH of 0.14% at 0.3 V
applied bias.[31,47] However, the only previously
reported H2ase photocathode in a tandem cell employed a
p-Si photoabsorber and achieved a STH of 0.006% for bias-free water
splitting.[25] The unassisted solar-to-fuel
conversion of the BiVO4||PVK|TiO2|H2ase tandem device was also more efficient than previous bacterial
biohybrid systems.[48] The PVK-H2ase system presented here shows superior performance to equivalent
earth-abundant molecular artificial and biological catalyst systems
reported to date.
Table 1
Solar-to-Fuel Efficiencies of State-of-the-Art
Tandem Devices that Employ Immobilized Earth-Abundant Molecular H2 Catalysts, a Bacterial Catalyst, and an Analogous Pt Device
System
Tandem Cella
Solar-to-Fuel/%
Product
Ref
platinum
BiVO4∥PVK|Pt
0.35
H2
(15)
synthetic
Ru|OD|TiO2∥NiO|OD|Co
0.05
H2
(42)
TaON|CoOx∥CuGaO2|OD|Co
0.0054
H2
(47)
enzymatic
IO-TiO2|OD|POs-PSII∥IO-ITO|H2ase
0.14 (0.3 V bias)
H2
(31)
BiVO4∥p-Si|IO-TiO2|H2ase
0.006
H2
(25)
BiVO4∥PVK|IO-TiO2|H2ase
1.1
H2
this work
bacterial
TiO2∥Si|TiO2|S. Ovata
0.38
acetate
(48)
OD = organic dye. See Table S2 for details.
OD = organic dye. See Table S2 for details.In conclusion, the combination of a biocatalyst with
a moisture-sensitive
perovskite photoabsorber has been accomplished, and this biomaterial
hybrid has subsequently been employed in overall tandem solar water
splitting. The perovskite–H2ase photocathode was
realized by (i) encapsulating the perovskite using a eutectic alloy,
metal foil, and epoxy resin and (ii) integrating the enzyme into a
hierarchical IO-TiO2 scaffold. The PVK|IO-TiO2|H2ase system achieved benchmark performance for photocathodes
driven by earth-abundant catalysts with a current density of −5
mA cm–2 at 0.0 V vs RHE, a positive onset potential
of +0.8 V vs RHE, a H2 production yield of 258 ± 55
μmolH cm–2 and a H2ase-based TONH of 1.9 × 106. A bias-free semiartificial water splitting device was produced
using the PVK|IO-TiO2|H2ase photocathode and
a water oxidizing BiVO4 photoanode. In a single-compartment
“leaf” configuration, the tandem PEC system was shown
to have an onset potential of −0.6 V and a solar-to-hydrogen
efficiency of 1.1% without applied bias. This work provides a new
benchmark for photocathodes and tandem PEC devices employing earth-abundant
molecular H2 production catalysts. The hybrid system demonstrates
the potential for bias-free fuel production and establishes perovskites
as a suitable photoelectrode material for the integration of biological
catalysts.
Authors: Fusheng Li; Ke Fan; Bo Xu; Erik Gabrielsson; Quentin Daniel; Lin Li; Licheng Sun Journal: J Am Chem Soc Date: 2015-07-13 Impact factor: 15.419
Authors: Chong Liu; Joseph J Gallagher; Kelsey K Sakimoto; Eva M Nichols; Christopher J Chang; Michelle C Y Chang; Peidong Yang Journal: Nano Lett Date: 2015-04-07 Impact factor: 11.189
Authors: Yixin Zhao; Nicholas C Anderson; Michael W Ratzloff; David W Mulder; Kai Zhu; John A Turner; Nathan R Neale; Paul W King; Howard M Branz Journal: ACS Appl Mater Interfaces Date: 2016-06-02 Impact factor: 9.229
Authors: Micaela Crespo-Quesada; Luis M Pazos-Outón; Julien Warnan; Moritz F Kuehnel; Richard H Friend; Erwin Reisner Journal: Nat Commun Date: 2016-09-06 Impact factor: 14.919
Authors: Dong Heon Nam; Jenny Z Zhang; Virgil Andrei; Nikolay Kornienko; Nina Heidary; Andreas Wagner; Kenichi Nakanishi; Katarzyna P Sokol; Barnaby Slater; Ingo Zebger; Stephan Hofmann; Juan C Fontecilla-Camps; Chan Beum Park; Erwin Reisner Journal: Angew Chem Int Ed Engl Date: 2018-07-17 Impact factor: 15.336
Authors: Vivek M Badiani; Samuel J Cobb; Andreas Wagner; Ana Rita Oliveira; Sónia Zacarias; Inês A C Pereira; Erwin Reisner Journal: ACS Catal Date: 2022-01-20 Impact factor: 13.700
Authors: Samuel J Cobb; Vivek M Badiani; Azim M Dharani; Andreas Wagner; Sónia Zacarias; Ana Rita Oliveira; Inês A C Pereira; Erwin Reisner Journal: Nat Chem Date: 2022-02-28 Impact factor: 24.274
Authors: Esther Edwardes Moore; Samuel J Cobb; Ana Margarida Coito; Ana Rita Oliveira; Inês A C Pereira; Erwin Reisner Journal: Proc Natl Acad Sci U S A Date: 2022-01-25 Impact factor: 12.779