Artificial Photosynthesis: Learning from Nature

Artificial photosynthesis has been devised and investigated in pursuit of solving 21th century's energy problem. Despite such advances in recent decades, applying the technology in real life is still a challenging subject for the scientists. As the term "artificial photosynthesis" stems from mimicking the natural photosynthesis, we can learn from the nature's strategies which have been evolved for 3.4 billion years. This review highlights important strategies of natural photosynthesis which can be borrowed for highly efficient and robust artificial photosystem for solar fuel production. Starting with a brief description of photosystem II in natural photosynthetic autotrophs, three relevant bioinspired strategies are discussed in this article: i) accumulative charge transfer, ii) photoprotection, and iii) self-healing. Then development of artificial photosystems mimicking those strategies will be discussed. Finally, remaining challenges and perspectives for future development of artificial photosynthesis are discussed.


Introduction
Photosynthetic autotrophs have been converting photon energy to chemical energy for 3.4 billion years;s ome of the convertede nergy was deposited on earth as forms of fossil fuels. Currently,h uman beings are consuming 13 TW of energy, in which8 1% is based on combustion of fossil fuels. [1] Along with the fast depletion of fossil fuels, the use of carbon-based fuels inevitably produces air pollutiona nd greenhouse gases. To solve this conflict between energy demanda nd ecological issues, sustainable and clean technologiesf or energy production are required. Sun provides1 71 000 TW energy on the surface of the earth, thus utilizingo nly 0.01 %o fi tw ill be just enough to meet the whole need of energy by human beings. In this context,d eveloping as ystem which can convert the photon energy to chemical energy is of great importance for the solar flux utilization.
Naturalp hotosynthesis utilizes solar energy to convert water and/or carbon dioxide into higher energy components such as carbohydrates. Artificial photosynthesis replicates the photochemicalp rocess of natural photosynthesis, however,i ti s more dedicated to the productiono fu seful fuels or valuable chemicals such as molecularh ydrogen (H 2 ), methane,m ethanol, etc. The first idea of artificial photosynthesis dates back to the beginningo f1 900s when GiacomoC iamician commented in his paper that human should shift from consuming fossil fuels to generating sustainable energy from the sun. [2] The first experimental demonstration of light-driven water splitting was so called "Honda-Fujishima effect". [3,4] The photoelectrochemical cell (PEC) was composed of TiO 2 photoanodea nd platinum black cathode. By illumination of the photoanode (l > 400 nm), O 2 and H 2 were generated at the photoanode and cathode, respectively.T he same group reported ap hotocatalytic CO 2 reductionw ith aqueous suspension of variouss emiconductor particles. [5] Since the pioneeringw orks by Honda and Fujishima, enormouse fforts have been made in pursuit of efficient, stable and cost-effective artificial photosystemsf or water splitting and CO 2 reduction.  Nature has developed severals trategies to run the photochemicalp rocesses of photosynthesis in an efficient and robust way:i )accumulativec harge transfer to run the multielectron process of product production efficiently, ii) photoprotection to preventt he system from being damaged by excess photon flux, and iii) self-healing to repair or replaced amaged proteins prolong the activity of whole system.T hose strategies have been developedt ob alance the maximum efficiency and sustainability upon fluctuating environmental conditions. Thus, closely mimicking these nature's strategies is ac hallenging but promising approacht orealize the efficient and stable artificial photosynthesis in real life application.T his approach has received recent attention, and many research groups are currently developing molecular components which mimic the natural photosystem for solar fuel production.
In this contribution, we would illustrate recentp rogresses on molecular systems for artificial photosynthesis by mimicking the nature's strategies. There is another research area of biohybrid systems directly using the part of nature, i.e.,b acteria or enzymes,which will not be discussed in this paper.

Overview of Photosystem II (PSII)
Production of organic matter andm aintenance of most forms of life on Earth is maintainedb yo xygenic photosynthesis process. This process converts light energy from sun into chemical energy,w hich is stored as highly reduced organic compounds. Twom ain photosynthetic reactionc enters (RC) namely,p hotosystem I( PSI) and PSII, are responsible for linear electron transport in oxygenic photosynthesis. In early 20th century Emerson and Arnold reportedt hat hundreds of chlorophyll (Chl) molecules operated together to produce one oxygen molecule. [50] This network of Chl molecules, which supplied excitation energy to photochemical catalytic center,w as later on named as "photosynthetic unit". In 1960 Hill and Bendall proposed that at wo-pigments ystem workedi ns eries in order to realize oxygenic photosynthesis where short-wavelength reaction and al ong wavelength reaction take place. [51] This hypothesis of Hill and Bendall, called zigzag model or "Z-Scheme" was experimentally provedb ys everal studies in 1961 and the terms PSI and PSII were coined. PSII, whichi samulti-subunit membrane protein that has the unique ability to extract electrons from water to result in molecularo xygen (O 2 )u pon reduction of Artificialp hotosynthesis hasb een devised and investigated in pursuit of solving the 21st century's energy problem. Despite significant advances in recent decades, applying the technology in real life is still ac hallenging subject for scientists. As the term "artificialp hotosynthesis" stems from mimicking natural photosynthesis, we can learn from nature's strategies which have evolvedo ver 3.4 billion years. This Review highlightsi mportant strategieso fn atural photosynthesis which can be borrowed for highly efficient and robust artificial photosystems for solar fuelp roduction. Starting with ab rief description of photosystem II in natural photosynthetic autotrophs,three relevant bioinspired strategies are discussed in this article:i )accumulative charge transfer,i i) photoprotection, and iii)self-healing. Next, development of artificial photosystems mimicking those strategies will be discussed.F inally,r emaining challenges and perspectives for future development of artificial photosynthesis are described.
plastoquinone. PSII constitutes at least 25 different protein subunits many of which are bound to the thylakoid membrane. [52,53] However,m inimal reaction center responsible for O 2 evolution has fewer members.
The most important constituents of the minimal reaction center of PSII are proteins D1, D2, a and b subunits of Cyt b 559 and PsbI subunit. This assembly of five is responsible for the charge separation. [53] Among these D1 and D2 are of great importance. D1 andD 2a re responsible for keeping all cofactors in as pecific positions ot hat the reactions can proceed in the right order. [54] The process starts with capturing of light energy by pigments within the antenna system.T hen the energy is transferredt ot he Cholorophylls (Chls) that involvest he primary electron donor,P 680, to generate an excited singlet state, P680*. Excited state relaxes back by reducing an earby Pheophytin (Pheo), am olecule similart oC hl in structure but missing the central Mg ion, forming the primary radicalp air P680 + Pheo À .R educed Pheophytin,P heo À ,t hen oxidized back to Pheo by giving an electron to the primary plastoquinone electron acceptor,Q A ,a ttached to the D2 protein. The electron is then transferred to the secondary plastoquinone electron acceptor,Q B ,b ound to D1. Q B is then reduced once more and gets protonated to yield the plastoquinol, Q B H 2 .A fter leaving the PSII system, this molecule is replaced by an oxidized plastoquinone molecule from ap ool of plastoquinone molecules located in the lipid bilayer.P 680 + ,t he oxidizedp rimary donor, is reduced by ar edox-activet yrosine residue, Tyr161 of the D1 protein. Tyr161 is afterwards reduced by the Mn 4 Ca cluster,c o-ordinated by aspartate, glutamate, and histidine residues in the D1 protein, including the C-terminus of matureD 1, and by residual glutamate of CP43. The Mn 4 Ca cluster exists in fivedistinct oxidation or Ss tates termed S 0 -S 4 ,w here the subscript indicatest he number of accumulated oxidizing equivalents. One molecule of oxygen occurs from two moles of water after formationo ft he S 4 state, resetting the enzymet ot he S 0 state, and so occurs with aperiodicity of four.

Accumulative Charge Transfer
Water splitting and CO 2 reduction are multi-electron processes (Scheme 1), which implies that multiple electrons should be accumulated at the catalytic center to runt he photochemical cycles. Even for the simplest processes for H 2 production or CO formation,i tr equires two electrons accumulated to ac atalyst. It should be noted that in general, photon absorption is a monoelectronic process, results in single charges eparation. In this context, it is required to accumulate multiple electrons or holes with severalr ounds of successive single photon absorption before chargerecombination occurs.
In PSII apparatus, multiple rounds of single electronic absorptiona re successfully coupledw ith extractiono ff our electrons from the Mn 4 Ca clustert or un water oxidation. The mechanism of water oxidation by PSII hasb een extensively investigated so far. [55][56][57] Schematic representation of the key cofactors in PSII is shown in Figure 1a.T he photolytic cycles tarts with light absorption of the chlorophyll pigment (P680) to form ap hotoexcited state (P680*), followed by charges epara-Dong Ryeol Whang received aPhD in materials science and engineering from Seoul National University in 2014. Dr.Whang is currently an assistant professor at Johannes Kepler University Linz, Austria. His research expertise and interests lie in the areas of artificial photosynthesis for solar fuel production. His research interests include catalysis for energy (fuel and chemicals) production combined with fine ligand synthesis, organometallic synthesis, density functional theory (DFT) based quantum chemical calculations, steady and time-resolved spectroscopy,a nd analytical chemistry.
Dogukan H. Apaydin obtained his BSc. degree in chemistry (2010) from Istanbul Te chnical University and his MSc. degree in polymer science and technology (2012) from Middle East Te chnical University in Turkey.Currently he is aPhD student/teaching assistant at Johannes Kepler University Linz in Austria focusing on (photo)electrochemical catalysis and electrochemical carbon capture.
Scheme1.Electron equivalents for water splitting and CO 2 reductionp rocesses. tion to generate radicalc ation (P680 + ). The hole can further be transferredt om ulti-electron-donor of tetramanganese cluster (CaMn 4 )t hrough tyrosine (Y Z )a sacharget ransfer mediator. It had been generally accepted that sequential extraction of four electrons from CaMn 4 is responsible for the catalytic cycle for water oxidation (Figure 1b). [58,59] Five-step reaction of CaMn 4 ,d enoted as S 0 through S 4 ,can be realized upon consecutive charget ransfer,w here S 0 is the ground state and S 4 is the four electron oxidized state. The S 4 state of CaMn 4 is quite unstablea nd rapidly reacts with water to produce O 2 .

Accumulation of Electrons
Wasielewski and co-workersf irst reported an accumulationo f two electrons in ad onor-acceptor-donort riad where two freebase porphyrin donors are attached to at wo-electron acceptor of N,N'-diphenyl-3,4,9,10-perylenebis(dicarboximide) (PBDCI) unit ( Figure 2a). [60] Upon flash photolysis of the porphyrin units with high laser intensities (up to 20 photons/molecule), the PBDCI unit was doubly reduced to give PBDCI 2À and two holes were stored in each porphyrin unites;o nly single charges eparation was observed with low laser intensities (< 50 photons/ molecule). Imahori and co-workers reported as imilar system where tetracyanoanthraquinodimethane (TCAQ)a nd zinc porphyrin were employed as acceptora nd units, respectively (Figure 2b). [61] Formation of doublyreduced TCAQ was successfully demonstrated using double pulse technique i) first pulse of 532 nm laser wasu sed to photogenerate the single charge separated state and ii)second pulse wasd elayedb y2 33 ps and set to 555 nm to selectively photoexcite the unexcited porphyrin units. Due to the distinct optical properties of singly-and doubly-reduced acceptor units, possible application of this molecular system as an opticals witch was proposed instead of artificial photosynthesis. However,p ioneered by these works, several examples of accumulative electron transfer in molecular systemsa iming on the multiple electrons utilization in artificial photosynthesis. [62][63][64][65][66] Af irst principle fort he accumulative electron transfer in APS is to store multiple electrons in one molecule to run the catalytic cycles.M olecular electron-reservoirs are compounds, which can store and transfer multi-electrons stoichiometrically or catalytically withoutd ecomposition or side reaction. [67][68][69][70] However,t he electron reservoir property in organometallic compounds is restricted to some classes of compoundst hat can withstand multiple redox changesw ithout molecular disruption.E specially for the transition metal complexes,i th as been noted that the population of do rbital upon photoexcitation or multiple reduction is am ain reason of the degradation of organometallic complexes.E lectron reservoirs can be designedw ith following strategies:i )the population of do rbital can be prevented if the orbitals affected by the electron addition are delocalizedo ver an extended ligand framework, and ii) tethering multiple electron acceptors which can store and stabilize excessive electrons in each electron reserving units.
The Park group recently reported aP t II water reduction catalyst (WRC) using the electron reservoir strategies. [71] Specially designed tetraphenylsilane (TPS) substituent was introduced to ad iimine ligand of am olecular Pt II WRC (TPSPtCl 2 in Figure 3). TPS has at etrahedral configuration where four phenyl rings are attached to ac entralS ia tom. The extended p-systems enable delocalization of excesse lectrons over the ligands, thus improvee lectron-reservoir character of the Pt II WRC. Electrochemicals tudy and density functional theory (DFT) based quantum chemical approachr evealed that TPS groups in TPSPtCl 2 create electron-reservoir characteristics andd ramatically enhanced electrochemical stabilityu pon two-electronr eduction. At urnover number of 510 000 was recorded for photocatalytic water reduction, which represents al arge improvement over the control complexes that do not contain the TPS substituents.
Sakai and co-workers reported as eries of Pt II WRCs which containsp yridinium or methyl viologen moieties as electron storing moieties. As imple Pt II terpyridyl complexw ith am ethyl pyridinium periphery,n amely PV 2 + ,w as synthesized and tested forl ight-driven H 2 production in combination with ethylenediaminetetraacetic acid( EDTA, YH 2 2À )a sasacrificial electron donor (Figure 4a). [72,73] It was found that PV 2 + formed a doubly-reduced species PV 0 by two consecutive photoinduced electron transfer (PET) steps, namely: 2À ] ! PV 0 + YH 2 À .T his molecular system well mimics the Z-Scheme,a lthough the two reductive quenching  Another class of Pt II polypyridine derivatives with viologen units was reported by the same research group (Figure 5a). The catalytic activity of the complex was enhanced by tethering viologen units as electron acceptors, which can temporarily collect the high-energye lectrons generatedd uring the photoinduced steps. In the case of the viologen-containing complexes, the ground state of the photosensitizing site, i.e., PtCl 2 (bpy) was immediatelyr egenerated from its reduced form after the first PETstep, thus restoringt he light absorption ability ( Figure 5b).As ar esult,t he Pt II bipyridyl complex with four viologen periphery, [PtCl 2 (5,5'-MV4)] 8 + + ,r ecorded one order of magnitude higher TON of 27 after 12 hofp hotolysis compared to PV 2 + + . [74] Further series study was carriedo ut and TONs of 35 and 25.2 were recorded with aP t II bipyridyl complex [75] anda Pt II terpyridyl complex, [76] respectively.W itht he same strategy, Ru II -based photosensitizers with peripheral methyl viologen units were designed to store and transfer multiple electrons to colloidal Pt WRC. Storing up to 8e lectrons in as ingle photosensitizer was achieved by tethering 12 methyl viologen units to a[Ru II (bpy) 3 ] 2 + center. [77,78] The Brewer group reported as eries of trinuclearc omplexes with am olecular structureo f[ (TL 2 Ru II BL) 2 M III X 2 ] 5 + (TL = terminal ligand,B L = bridging ligand,M = Ir or Rh, and X = halogen ligand). [79][80][81] Twop hotosensitizing Ru II units were bridged to the central electron accepting units to direct the accumulation of electrons. While Ir III center can store two electrons in itsl igands without affecting the oxidation state of the metal center, Rh III -basedc omplexesu ndergo metal-centered reduction to form Rh II species upon one electron reduction. Then Rh II intermediate disproportionates to form Rh III and Rh I species with loss of the two chloride ligands. The free coordination site of Rh I then opens ap ossibilityf or substrates such as protont oa nchor to form metal-hydride, followed by H 2 production. SeveralR h III -based trinuclear complexes were subjected to photocatalytic H 2 production [82][83][84][85] and TON of 1300 was recordedw ith [((Ph 2 phen) 2 Ru(dpp)) 2 RhBr 2 ] 5 + + (Ph2phen = 4,7diphenyl-1,10-phenanthroline and dpp = 2,3-bis(2-pyridyl)pyrazine, Figure 6a)i nc ombination with N,N'-dimethylaniline (DMA) as as acrificial electron donor. [85] Recently,C ollomb and co-workersd eveloped at rinuclear complex Ru 2 Rh,w here two Ru II photosensitizing units are linked to the Rh III center with a bridging ligand of 1,2-bis[4-(4'-methyl-2,2'-bipyridinyl)]ethane ( Figure 6b). [86] Non-conjugated ligand was used to minimize the electronic coupling between the photosensitizing and electron accepting units thus maintain the original physical proper-   ties of parent molecules,i .e., [Ru II (bpy) 3 ] 2 + and[ Rh III (bpy) 2 Cl 2 ] + . The authors highlighted ah igh TON of 430 in pure aqueous system,h owever,i th as been noted that the photocatalytic activities of the polynuclear systems have been rather low compared to those of multi-component systems. [87]

Accumulation of Holes
Te tramanganese clusteri sr esponsible for the accumulationo f four holes for O 2 evolution from water in PSII, vide supra. With this in mind, several photosensitizer/donor or photosensitizerdonor systemsh ave been made to mimic the chargea ccumulation, where the donors are based on mono-or polynuclear manganese complexes. Styring and co-workers reported either inter-and intra-molecular accumulation of holes in dinuclear Mn 2 II,II unit in combination with Ru II polypyridine moiety as a photosensitizer (Figure 7a). [88][89][90] In the presence of pentaamminechlorocobalt III chloride as an electron acceptor, photoexcitation of Ru II centerr esultedi nastepwise extractiono ft wo electrons from the Mn 2 II,II .E PR spectroscopy revealed continuous formationo fC o II species and Mn 2 III,IV species, indicating that three holes were accumulated in the dinuclearm anganese moiety (Figure 7b). [88] Collomb andD eronzier et al. reported similar systems where three Ru II tris(bipyridine) complexes are tethered to aM n II tris(bipyridine) center. [91,92] It wasp roposed in these studies that upon one electron reduction of the Mn II center,the manganese complex weredecomposeda nd irreversibly transformed into corresponding di-m-oxo bridged Mn 2 III,IV dimerics pecies (Figure 7c). Despite the presence of Mn II -based units, those systems did not show any catalytic activity of O 2 production from water.
Recently,d ye-sensitized inorganic/organic hybrid systems have been proposed to be an alternative approach for facilitating accumulative charget ransfer in molecular systems. Inor-ganic nanocrystals such as TiO 2 nanocrystals can utilized as an electron/hole reservoir.T herefore, hybridizing them with organic/organometallic dyes offers multiple charge separation after several rounds of single photon excitation. The Hammerstrçm group reportedaseries of inorganic/organometallic hybrids ystems where Ru II -polypyridine photosensitizers with oligotriarylamine (OTA) donors are linked to TiO 2 nanoparticles as an acceptor. [93,94] Careful and intensives tudyi nc ombination with double pulse experiments showede fficient accumulative electron transfer resulting in the formation of two-electron oxidized state of the OTAu nit upon successive excitationb yt wo photons. Among the systems studied, one with OTARu dye anchored onto nanocrystalline TiO 2 (Figure 8) showede fficient accumulative electron transfer upon successive photoexciation with double pulse experiments. Once the Ru II photosensitizer was photoexcited by first excitation pulse, the excitonw as in-  jected rapidly into the TiO 2 nanocrystalw ithin 1pst of orm Ru III species. Then the ground state of Ru II center was recovered by oxidizingO TA within 1nsr egime, resulting in TiO 2 (À) -Ru II -OTA + state. With second photoexcitation, one more round of charge separation occurred to yield doubly charges eparated state of TiO 2 (2À) -Ru II -OTA 2 + .T he overall yield of accumulative charge transfer was reported as high as 100 %.

Photoprotection
The intensity of the solar radiation fluctuates and whenever the light intensity is greater than that neededt os aturate the photosynthesis, the plant or algae faces ad anger of being damagedb yt he excesso fi ncoming photons. For example, when light absorption exceeds the capacity of photosystem in PSII, triplet excited state of chlorophyll ( 3 Chl) is generated either by intersystem crossingf rom its singlet excited state ( 1 Chl) or by charger ecombination of the primary radical pairs between PSII primary donor and pheophytin (P680 + /Phe À ). The 3 Chl is ap otent sensitizer for molecular oxygen forming singlet oxygen ( 1 O 2 )w hich can cause oxidatived amage to the pigments,l ipids and proteins of the photosynthetic system. To cope with fluctuating irradiance and prevent the cell or protein from getting damaged, photosynthetic organisms have evolveds everal photoprotective mechanisms:i )adjustment of light-harvesting antenna size, ii)thermald issipation of excess absorbed light energy,a nd iii)scavenging reactive oxygen species. [95][96][97][98][99][100][101][102][103][104][105][106][107][108][109][110] Ther egulation of excessp hotons thus can prolong the lifetime of the photosynthetic apparatus. Achieving selfregulating behavior in molecular system is ac ritical challenge to prolongt he artificial photosynthesis.

Adjustment of Light-Harvesting Antenna Size
One implication is that the light harvesting antennas in Chls can rearrange in their protein matrix upon either al ong term or as hort term fluctuation of light intensity. [111,112] During longterm acclimation to fluctuatingl ight intensities,c hanges in antenna size are due to changes in light-harvestingc hlorophyll protein complex( LHCs gene expression [113][114][115] and/or LHCs degradation. [116] In short term modulation of the antenna sizes, overexcitation of PSII resultsi nd etachment of LHCs from PSII, thus decreases the effective size of LHCs. [117] In both cases,t he overall light absorptiono fp hotosystem can be reversibly modulated with different light intensity. Molecular logic gate strategy can be borrowed to realize the light-intensity dependent regulation of photochemical processes. Am olecular logic gate is am olecule that performs al ogical operation based on physical or chemical inputs.O nce the logic gate is operated by means of photons, that is, photochromic molecule, we can modulate the photophysical properties of the molecule thus give ways to form ap hoton regulator.Either photoinduced intramolecular energy or electron transfer processes have been successfully controlled by using the photochromic units. Effenbergera nd co-workers developed aD -B-A system with ON/OFFp hotoswitchable intramolecular energy transfer by incorporationo faphotochromic fulgimide between anthracene and coumarin (Figure 9a). [118] Ts uchiyar eported an optical control of photo-induced electron transfer by incorporating two porphyrin units with ap hotoisomerizable azobenzene linker (Figure 9b). [119] Inspired by those works, significant efforts have been made on the development of artificial photoregulation systemso nc atalysts for chemicalr eactions, however,m osto ft he works have been focusing on the light-induced activation of the catalytic processes and downregulation of the catalytic activities are still scarce. [120] Although actual photoregulation has not been shown in those papers, they provided ag ood photochemical background foru sing those systems for photoregulation in artificial photosynthesis.
The molecular logic gate concept was successfully adapted to molecular photoregulatorys ystem.G ust and co-workers reported as eries of supramolecular systems which can downregulate the high intensity light by activation of by using thermally-reversible photochromic units which can reverse the photochemical process by thermale nergy (Figure 10). [121] In the ground state, the dye exists in the dihydroindolazine (DHI) form (1c) which can absorb only blue lights. When 1c absorbs light, it photoisomerizes into an open form of colored betaine (BT), depicted as 1o. Molecule 1o thermally converts back to 1c with at ime constant of 37 sa t2 5 8C. Based on this, the Figure 9. a) Chemical structure of coumarin isomersa nd conceptfor switching intramolecular energytransfer in amolecularsystem. [118] b) Chemical structures of electron-rich and electron-deficient porphyrins (upper) and process of photoswitching (lower).R eproduced with permission from Refs. [118] and [119]. photoregulatory system of naturalp hotosynthesis was successfully demonstrated by adjusting the absorption cross-section of the dye.U nder low light intensity conditions, the molecule mainly exists in ac losed form (1c), and undergoes photoinduced electron transfer from porphyrin to fullerenew ith a quantum yield of 0.82;a st he light intensity increases, formation of BT leads to quenching of the porphyrin excited state, reducing the quantum yield to as low as 0.27.

Carotenoids
Carotenoids are important pigments in photoregulatory system of photosynthetic bacteria. They can either prevent the formationofhighly destructive singlet oxygen by thermaldissipation of 1 Chl [122][123][124][125][126][127][128][129] or quenching the already produced singlet oxygen to its triplet ground state. [130] In higher plants, there are three carotenoid pigments that are active in the xanthophyll cycle:v iolaxanthin, antheraxanthin, and zeaxanthin ( Figure 11). Duringl ight stress, violaxanthin is converted to zeaxanthin via the intermediate antheraxanthin, which plays a direct photoprotective role acting as al ipid-protective anti-oxidant and by stimulating non-photochemical quenching within light-harvesting proteins.A nother role of carotenoids for photoprotective reactioni st hat they can directly quench already generated singlet oxygens. Duet ot he low-lying triplet excited state of the carotenoids, triplet carotene is formed as the singlet oxygen returnstoi ts ground state.
The carotenoid photoprotection was successfully mimicked with am olecular system where ac arotenoid moiety was covalently bondedt oaporphyrin (Figure 12 a). [131] Bleaching of diphenylisobenzofuran( DPBF), which is vulnerable to single oxygen, upon irradiating the aerated solution of the dyes were investigated. If the solutionc ontains only at etra-arylporphyrin sensitizer in addition to DPBF,s inglet oxygen is produced under illumination andr apidlyr eacts with DPBF,r esulting in the fast bleaching of the dye. In the presence of the dyes I or II,asignificant decrease in the rate of photodestruction of DPBF was observed because of the fast quenching of the triplet excited state of the porphyrin unit by the carotenoids, thus prevents the O 2 sensitization (Figure 12 b).

D1 Protein Repair Cycle in PSII
Despite the existence of photoprotective mechanismsi nP SII, photodegradation occurs even at av ery low irradiance. [132][133][134] Severald ifferent mechanisms have been proposed for the PSII photodegradation;r egardless of the mechanism,i ti sg enerally acceptedt hat the irreversible photodamage of D1 protein is responsible for the photoinhibition of PSII. [98,100,135] Once the D1 protein in PSII is irreversibly damaged by photons, am echanism is activated where the damaged part of the PSII complex are replaced by af resh one to restore the photochemical activity.Such as elf-healing mechanism is termed as "PSII repair cycle". Understanding the full cycles of self-healing is still underway,h owever,i th as been assumed to consist of an umber of distinct steps shown in Figure 13.
In 2008, the Nocera group reported ap reparation of Co IIbased electrocatalyst, namely Co-Pi,w hich is capable of performingw ater oxidation. [137] They observed depositiono ft he catalystu pon the oxidativep olarization of an inert indium tin oxide (ITO) electrode in Co II ionc ontaining phosphate buffer. The catalyst was formed in situ and operated at low overpotential of 280 mV.I ndeed, ap lausible self-healing pathway was proposed in the paper.T he catalytic system was further studied by the same group and the self-healing capability was confirmed. [138] Leaching of cobalt waso bserved during the water electrolysis, while by applying potential in the presenceo fa phosphate counter ion resultedi naredeposition of the catalyst. It was noteworthy that the phosphate electrolyte ensures then the stability of the catalytic film, by promoting redeposition of Co III ,g enerated from electrochemical oxidation of Co II ions leached in solution ( Figure 14). Av ery similars trategy with Co-Pi,M nO 2 water oxidation catalystw as reported by Najafpour et al. [139] Self-healing of aM nO 2 electrocatalytic film in the presence of Ce IV was demonstrated.
As an important step forward to artificial photosynthesis, Nocera's self-healing WOC was further applied for ap hotoanode visible-light driven water splitting photoelectrochemical cells (PECs). Gamelin andc o-workers reported aP EC where the photoanode wasp repared by electrodepositing Co-Pi on the surfaceo fp hoto-active hematite (a-Fe 2 O 3 )i nstead of using ITO as as ubstrate. [141][142][143] Under AM 1.5 solar irradiation, the PEC showed reduction of the bias voltage (> 350 mV) for water splitting. Steinmiller and Choi reported another photoelectrochemicala pproachb yp hotochemically depositing Co-Pi on ZnO rods. [144] By irradiating UV light instead of applying external bias, photogenerated holes in the ZnO rods were used to oxidize Co II ions to Co III ions to precipitate the Co-based catalyst ontot he surface. The photodeposition methode nsures the deposition of the catalyst where the photogenerated holes are most readily available, thus provides optimal catalytic site for catalysis. In the presence of Co-Pi,t he ZnO photoelectrode showedr eductiono fo nset potentialb y0 .23 Vw ith enhancement of anodic photocurrento bserved in aw ide potential range.I nb oth systems by Gamelin and Choi, the leaching of the catalytic specieswas ruled out due to the self-healing ability of the Co-Pi catalyst, thus prolonged the catalytic activity. Since the reports, Co-Pi has been coupled with various lightactive semiconductors to enhancet he catalytic activity of a photoanode for O 2 evolution, implying the importance of selfhealing for prolonged and efficient catalysis. [145][146][147][148][149][150][151][152][153] By incorporating Co-Pi with at riple junction amorphous silicon (3jn-a-Si) photovoltaic cell and H 2 -evolving NiMoZn ternary alloy,avery high efficient solar-water-splitting device,n amely "artificiall eaf" was constructed with ad evice configuration of Co-Pi/3jn-a-Si/NiMoZn (Figure15). [154,155] Overall solar-to-fuels efficiencies (SFE) of 4.7 %w as recorded with the artificial leaf, corresponds to the overall water splitting efficiency of 60 %.
The Bocarsly group reported ap hotoelectrochemicals plitting of water in an ambient condition using polycrystalline CuRhO 2 as ap hotocathode. [156] It was proposed that the presence of O 2 in the electrolyte solution either prevents the formationo fm etallic Cu or rapidly reacts with trace quantities of Cu to regenerate CuRhO 2 .
One implication on the organometallic catalysts is that they are pronetoundergo ligand dissociation in the photolysis conditions. Eisenberg and co-workers showed regeneration of the ligand-dissociated catalysts by adding free ligand that reconstituted the catalyst. [157,158] Photocatalytic H 2 evolution kinetics of ac obalt catalyst, Co III (dmgH) 2 pyCl (whered mgH = dimethylglyoximate and py = pyridine) were compared with various concentrations of free dmgH 2 (Figure16a). In the absence of the free ligand,t he H 2 evolutionc eased in 5h with at urnover Figure 13. Schematic representation of PSII repair cycle.R eproduced with permission from Ref. [136].C opyright (2011) The Royal SocietyofC hemistry. Figure 14. Schematic representation of self-healing process in oxygen-evolving Co III catalyst. Reproduced with permissionfrom Ref. [140]. Copyright (2009) The Royal SocietyofC hemistry. [140] ChemPhotoChem 2018, 2,1 48 -160 www.chemphotochem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim 156 Reviews number of 360, while in the presence of 12 equiv of dmgH 2 , the catalytic activity prolonged 12 hr ecording at urnover number of 900 (Figure 16 b). On the other hand, the addition of free ligand sometimes deteriorates catalytic activity when a metallicc atalystw as used. The Bernhardg roup reported a photocatalytic H 2 production using Ir III (ppy) 2 (bpy) (where ppy = 2-phenylpyridine and bpy = 2,2'-bipyridine) photosensitizers and colloidal Pt catalyst. [159] Addition of 50 equiv free bpy ligand to the system lowered the H 2 production activity to 40 %o fi ts originala ctivity instead of regeneratingt he Ir III photosensitizer,w hichw as attributed to poisoningo ft he Pt catalyst.

Summary and Outlook
The purpose of this Review has been to understand and benchmark the natural photosynthetic system andd escribe their incorporation into artificial photosynthesis. Towards this end, three important and unique strategies in PS II were stated:i )accumulative charget ransfer,i i) photoprotection, and iii)self-healing.T he natural photosynthetic system has been evolvedt of ulfill the strategies that the active units have optimal spatial positions in ap rotein matrix for the balance between efficient chargea ccumulationa nd photoprotective operation. Additionally,t he damaged proteins can be replaced throughas elf-healing mechanism to furtherp rolongt he photosynthetic activity.M any challenges have been made as analoguesf or each strategy for artificial photosynthesis. Successful   www.chemphotochem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim achievements have been made using either supramolecular chemistry or an organic-inorganic hybrid approach. It was shown also in inorganic systems, that is, Nocera's Co-Pi catalyst, that the microscopica pproacho nt he molecular or atomic scale wasvery useful to controlt he catalytic activity.
The promisingr esults reported so far provide ag ood blueprint for artificial photosynthesis;t his research field requires furtherd evelopment for real life application. The current status in artificial photosynthesisi st hat each component responsible for light harvesting, charge transfer and accumulation, and catalysis can be optimizedind ifferent working conditions, however,c ombining them in as ingle device as ac omplete module is still ac hallenging task. Careful orchestration of the three strategies mentionedi nt his article in one system is required. In the integrated systemw eh avet ot ake accounto fc riticalf actors: 1) Balancing between chargeaccumulation and the regulatory mechanism to aim to "kill two birds with one stone", that is, keeping optimal solar fuel production efficiency while the stabilityi sg uaranteed under fluctuating solari rradiation; 2) Self-healing should work while the catalytic system is under operation; 3) The self-healing mechanism itself should be self consistent, not by stoichiometrically consuming other reagents or additives.
Supramolecular chemistry can take ap art to mimict he role of protein matrix in naturalp hotosystem. Controlling charge separation, -transfer,a nd accumulationa long with self-healing can be accomplished with as upramolecular approach. Combining the benefits of homogeneous and heterogeneous system can also provide as olution. The molecular system has merits of synthetic variety,e ase of physical and chemical property tuning,h igh catalytic site accessibility,w hile the disadvantages of low stability and low conductivity can be solved by complementary inorganic systems. The multidisciplinary approach is still challenging, but we can learn from the 3.4 billion years of evolution in nature.