A Biologically-Inspired Electrochemical Half-Cell for Light-Driven Generation of Hydrogen

By: Rebecca A. Grimme1 and John H. Golbeck*1,2, Material Matters 2008, 3.4, 78.

1Department of Chemistry, The Pennsylvania State University
University Park, PA 16802
2Department of Biochemistry and Molecular Biology,
The Pennsylvania State University, University Park, PA 16802
*E-mail: jhg5@psu.edu


An economy based on hydrogen has long been touted as one of the most promising routes to a sustainable energy future. Hydrogen gas (H2) is a particularly clean source of energy that can be utilized efficiently in hydrogen-based fuel cells. Because the product of energy conversion is H2O, the use of H2 does not exacerbate the already existing problem of CO2 accumulation that is a major component of global climate change. Despite its desirable attributes, H2 is not found in nature and methods must be employed to generate it using other sources of energy.

Commercial methods of producing hydrogen are expensive. These processes include electrolysis of water and steam reforming of methane. The generation of hydrogen from the electrolysis of water takes place in an electrolytic cell where an electrical power source is connected to a cathode and to an anode, each in separate vessels connected via a salt bridge. The negatively charged electrode combines electrons with protons, and this is where hydrogen is produced. The electrolysis of water requires a larger input of energy than would otherwise be necessary due to the fact that 100% of the electrical power put into the cell is not converted into the chemical bond energy of hydrogen.1 Steam reforming of methane is less expensive than the electrolysis of water and is therefore the preferred industrial method of production. The heating of steam and methane to high temperatures in the presence of a nickel catalyst results in the generation of H2 and CO. The further reaction of H2O and CO produces additional H2 and CO2. The by-product of the reaction, CO2, however, adds to the problem of climate change. Also, because energy is lost by the conversion of methane into hydrogen, it makes little sense to use one perfectly good fuel to generate another.

With all of the drawbacks associated with currently utilized methods of hydrogen production, researchers are increasingly looking to sunlight as a means of supplying the necessary energy input. Sunlight is abundant, widely distributed, and virtually inexhaustible. The amount of solar energy that strikes the surface of the earth in one hour is more than sufficient to satisfy the total global demand for energy in an entire year.2 Hence, the photocatalytic production of hydrogen is emerging as an attractive route to the generation of a renewable source of fuel.

back to top Back to Top

Emerging Technologies

A large amount of research has focused on the generation of H2 by systems that employ gold (Au, Aldrich Prod. No. 636347) or platinum (Pt, Aldrich Prod. No. 685453) nanoparticles supported on semiconductor surfaces such as TiO2 (Aldrich Prod. Nos. 635057, 635049, 635065). These semiconductor systems function by absorbing a photon and creating an electron-hole pair. The electron migrates to the Au or Pt nanoparticles that are located on the semiconductor surface. These nanoparticles catalyze hydrogen production. A major drawback to this type of system lies in the fact that the band gap of TiO2 semiconductors is large, and photons with sufficient energy to generate the electron-hole pair can be found only in ultraviolet (UV) radiation.6 UV, which is light with wavelengths between 200 and 400 nm, constitutes only a small percentage of the solar radiation that reaches the surface of the earth. In an attempt to overcome the band-gap limitation, organic dyes can be introduced into these semiconductor systems to increase the absorption of solar radiation.3-5 These dyes absorb photons in the visible portion of solar radiation and are thus able to inject electrons into the conduction band of the semiconductor materials.4,5 Once electrons are in the conduction band, they migrate to the Au or Pt nanoparticles loaded on the surface, thereby enabling hydrogen production. While these dye sensitized systems increase the functional working wavelengths that are capable of sustaining the charge-separated state necessary for hydrogen production, organic dyes are degraded by illumination and moisture, limiting the usefulness of these devices.

back to top Back to Top

Biological and Biohybrid Systems for Light Driven Hydrogen Production

While the dye-sensitized and semiconductor-supported nanoparticle systems hold promise, newer investigations have focused on using both biological systems and biohybrid designs to photocatalytically generate H2. It is generally accepted that any photocatalytic system for generating a usable fuel from sunlight will require three components: a module that converts sunlight into an electric current, a module that catalyzes the reduction of protons to H2, and a linker that facilitates the transfer of electrons from the light module to the catalytic module.

One such system couples Photosystem I (PS I), a naturally occurring light harvesting complex, to hydrogen evolving catalysts. The advantages of utilizing PS I as the energy-converting module are discussed in detail below. In vivo experiments using photosynthetic organisms have shown that hydrogen can be evolved by the enzymes nitrogenase (N2ase) and hydrogenase (H2ase).7,8 Hydrogen is produced as a byproduct of nitrogen fixation by N2ase and it is a natural product of H2ase when an electron donor is present.7 When the reducing equivalents from PS I are directed to H2ases or N2ases (the linker being a redox protein), cyanobacteria and microalgae evolve hydrogen.8 A major drawback to these in vivo systems is that the efficiency of hydrogen generation will always be limited due to the amount of solar energy necessary to sustain the growth and metabolism of the organism. A further drawback is that N2ases and H2ases are oxygen sensitive, making hydrogen generation problematic in the presence of oxygenic photosynthesis.

In vitro systems that combine biological and/or non-biological components in novel ways were the first attempt at a hybrid approach for hydrogen production. These in vitro systems utilize isolated proteins or proteins within chloroplasts in conjunction with metal deposition and/or naturally occurring H2ases. A PS I-H2ase construct has been created by fusing the gene encoding for the [NiFe]-H2ase from Ralstonia eutropha H16 with the gene encoding the PsaE (PS I stromal) protein. When the H2ase/PsaE fusion product is re-bound to a PsaE deletion mutant of PS I, hydrogen is evolved, albeit at low rates (0.2 μmol H2 mg Chl-1 h-1).9 Greenbaum and coworkers have pioneered the use of PS I/metal nanoparticle constructs for the photocatalysis of H2. In this work, Pt and other metals have been directly photoprecipitated on the stromal side of both isolated PS I proteins and PS I contained within spinach chloroplasts.10,11 Illumination of these platinized chloroplasts and PS I proteins enables the production of H2, but again at low rates of 0.2 to 2.0 μmol H2 mg Chl-1 h-1. In a more recent study, the cross-linking of the naturally-occurring electron donor plastocyanin was shown to double the rates of hydrogen production achieved from platinized PS I.12

back to top Back to Top

Efficient Photon Capture and Energy Conversion by PS I

PS I is a light harvesting complex that is located in the photosynthetic membranes of plants and cyanobacteria, a photosynthetically active bacteria. The major purpose of PS I is to use the energy of light to transfer electrons from high potential (i.e. low energy) redox proteins across the membrane to low potential (i.e. high energy) redox proteins.13 Figure 1 depicts the arrangement of PS I proteins within the membrane. Although PS I comprises 13 proteins, only PsaA, PsaB, and PsaC are of interest for this article. PsaA and PsaB are intramembrane proteins which support the core electron transfer cofactors of PS I, while PsaC lies outside of the membrane and acts as an interface to shuttle electrons from within the membrane to soluble low potential electron accepting proteins. Figure 2 affords a more detailed look at the organization of PS I. Additional intramembrane proteins surround the PsaA/PsaB heterodimer and support ~100 antenna chlorophyll molecules that are active in light harvesting. 14 These antenna pigments in cyanobacterial PS I are chlorophyll a (Chl a) molecules that are capable of absorbing photons with wavelengths shorter than 700 nm. This absorbance corresponds to 43-46% of the total solar radiation that reaches the surface of the earth. 15 When a Chl a molecule absorbs a photon, an excited state is created; the energy is ultimately transferred by resonance energy transfer to the primary electron donor of PS I, a Chl a special pair, termed P700. The arrangement of the core electron transfer cofactors is shown in Figure 3. When the exciton reaches P700, a charge-separated state occurs between P700 and the primary electron acceptor, A0, another Chl a molecule. Ultimately, the electron is transferred through the other electron transfer cofactors to FB, the terminal cofactor within PS I. The FB cluster has a midpoint potential of -580 mV, which is more than sufficient to reduce protons to H2.13 The quantum yield of PS I approaches 1.0 which means that nearly all of the photons that PS I absorbs are converted to the charge separated state P700+/FB -.

Figure 1 This cartoon depicts Photosystem I within the thylakoid membrane. PsaA and PsaB (light green) are an intramembrane protein heterodimer that contain the core electron transfer cofactors within PS I—P700, A0, A1, and FX. PsaC lays outside of the membrane on the stromal side of PS I and harbors the electron transfer cofactors FA and FB. Additional intramembrane proteins (dark green) support antenna chlorophyll molecules that are responsible for light harvesting. High potential (low energy) electron donating proteins (plastocyanin or cytochrome c6) provide electrons to P700. The light-driven electron transfer through the membrane is photocatalyzed by photons absorbed by the antenna chlorophyll molecules. Electrons ultimately arrive on the stromal side of PS I at FB and can then transfer to soluble low potential (high energy) electron accepting proteins (ferredoxin or flavodoxin).

Figure 2 a) The intramembrane α-helices of PS I (yellow tubes) act as scaffolds for antenna Chl a molecules (green) and the core electron transfer cofactors; the [4Fe-4S] clusters, FA and FB, are visible on the stromal side. b) Rotating structure (a) 180° and removing the protein scaffold affords a look (top down from the stromal side) at the organization of the antenna Chl a molecules around the core electron transfer cofactors (circled in grey).

Figure 3 The core electron transfer cofactors are arranged to allow for light-induced electron transfer to occur rapidly from P700, through the cofactors, to FB. A charge-separated state is first established between P700 and A0. The electron is then transferred to A1, a bound phyloquinone molecule, and then to three [4Fe-4S] clusters. The first of these, the inter-polypeptide [4Fe-4S] cluster, FX, is ligated by cysteine residues provided by both PsaA and PsaB. The stromal protein PsaC harbors the two terminal [4Fe-4S] clusters, FA and FB.

Electron transfer on the PS I acceptor side is thermodynamically favorable, as the midpoint potential of each of the subsequent cofactors is more positive than the previous one. Figure 4 shows the potentials of the electron transfer cofactors as a function of their distance from P700 as well as forward electron transfer and charge-recombination times. The electron transfer from P700 to FB is rapid (~200 ns) and the lifetime of the charge-separated state, P700+/FB -, is long (~65 ms).16 Provided the electron is transferred away from the FB cluster within this lifetime, charge recombination will not occur and electron can be harnessed for useful work. In the case of normal photosynthesis, this work is the reduction of ferredoxin or flavodoxin (and the subsequent production of NADPH), but if the electron can be removed at FB - directly, it can be used to reduce protons to H2.

Figure 4 As the electron is transferred away from P700, the potential of the cofactors becomes more positive, thus the transfer is thermodynamically downhill and favorable. Further electron transfer from FB – to a H2 evolving entity (Pt nanoparticle or H2ase enzyme) is also thermodynamically favorable as the redox potential for H2 evolution is thermodynamically more positive than that of the FB cluster.

back to top Back to Top

Hydrogen Catalysis by Pt Nanoparticles

As in the case of semiconductor systems, Pt nanoparticles can be used as the catalyst for hydrogen generation. H+ ions adsorb onto the Pt surface and combine with an electron to form an H atom. A covalent bond is catalytically generated between two adsorbed H atoms to yield H2. H2 then desorbs from the Pt surface. At pH 7.0, this process occurs at a midpoint potential of -420 mV. If the electrons are derived from the FB cluster of PS I, there is ~160 mV of thermodynamic driving force for this reaction to occur. This translates to an equilibrium constant >102, indicating that this reaction will be a highly favorable reaction on thermodynamic grounds.

In practice, a system could be set up in which PS I and Pt nanoparticles are suspended in solution, however hydrogen generation would most likely be of low yield due to the fact that the interactions between nanoparticles and PS I would be controlled by slow diffusion chemistry. The speed of diffusion decreases as size of the body in motion increases. In this case, both the PS I and the Pt nanoparticles are large entities and diffusion would likely be too slow to transfer the electron from FB - to the nanoparticle surface before the charge recombination between P700+ and FB - would occur. In order to avoid this inevitable loss of energy, a direct link should be made between the PS I and the Pt nanoparticle.

back to top Back to Top

Molecular Wires Form a Covalent Pathway

Molecular wires are the answer to the diffusional limitation in electron transfer. A molecular wire, in the form of an aliphatic or aromatic hydrocarbon chain, can be used to connect PS I with the Pt nanoparticle. On one hand, the molecular wire should be sufficiently long to shield the protein from the nanoparticle surface to limit protein denaturation. On the other hand, the molecular wire should be sufficiently short enough to allow for efficient energy transfer between the two modules. Because the charge-separated state P700+/FB - is stable for ~65 ms, efficient electron transfer away from FB must occur on the order of 1 ms. Marcus theory, which relates the rate of electron transfer to the distance between the cofactors (as well as the Gibbs free energy change, reorganization energy, and temperature), governs the maximum distance between PS I and the Pt nanoparticle for optimal electron transfer. Under ideal conditions, for the electron to be transferred on the microsecond timescale, the distance between the cofactors should be shorter than 2.0 nm. Short aliphatic and aromatic hydrocarbons with thiol moieties easily functionalize Pt nanoparticle surfaces and are commercially available. Unfortunately, a direct bond cannot be made to native PS I without modification.

back to top Back to Top

Manipulation of PS I

In order to attach the molecular wire, a variant of the PsaC subunit must be engineered in which the native cysteine ligand to the FB cluster is replaced with a glycine residue. Glycine lacks a side chain, and is hence incapable of ligating the Fe atom. However, its use opens up a site to which a molecular wire can be attached. The iron-free form of the PsaC variant is expressed in Escherichia coli and the FA and FB clusters are inserted using inorganic [4Fe-4S] clusters ligated by 2-mercaptoethanol (Sigma Prod. No. M7522). The PsaC variant can then be re-bound to previously-prepared PS I cores that lack the PsaC subunit. Analytical techniques show that two [4Fe-4S] clusters are inserted into the protein despite the absence of one of the cysteine ligands to FB. This anomaly has been explained by the fact that the [4Fe-4S] clusters are inserted into the protein in vitro by a ligand exchange mechanism in which the 2-mercaptoethanol ligands are displaced by the protein cysteine ligands. At the glycine position, the 2-mercaptoethanol is retained, where it functions as a so-called rescue ligand for the FB cluster. The insertion process, incidentally, is driven to completion by the entropic gain realized when seven 2-mercaptoethanol molecules are released into solution during the insertion process. Thiol functionalities in the form of a molecular wire can then readily displace the single 2-mercaptoethanol ligand through facile sulfur-iron displacement reactions.17,18

back to top Back to Top

Catalytic Hydrogen Production Using Bioconjugates

Bioconjugates composed of rebuilt PS I containing the glycine variant of PsaC, 1,6-hexanedithiol (Aldrich Prod. No. H12005) molecular wires, and 3 nm Pt nanoparticles are assembled as shown in Figure 5. Sodium ascorbate (Sigma Prod. No. A7631) serves as the sacrificial electron donor to dicholorophenylindophenol (DCPIP) (Sigma-Aldrich Prod. No. 119814), which donates electrons directly to P700+. When the module is illuminated, H2 is generated at a rate of 9.6 μmol H2 mg Chl-1 h-1 (0.23 mol H2 mol PS I-1 s-1) over a period of 24 hr. The addition of cytochrome c6, which is a superior electron donor to P700+, results in hydrogen generation at a rate of 49.3 μmol H2 mg Chl-1 h-1 (1.17 mol H2 mol PS I-1 s-1).19

Figure 5 This bioconjugate is composed of Photosystem I rebuilt from recombinantly expressed C13G/C33S variant PsaC, PsaD, and P700/FX cores, a 1,6-hexanedithiol (Aldrich Prod. No. H12005) molecular wire, and either Au or Pt nanoparticles. The molecular wires serves not only as a covalent link between PS I and the nanoparticle but also acts as a conduit for electron transfer. The electron donor, cytochrome c6, is not depicted here.

While the initial rates for this system were already promising, continued research has yielded better performance. Altering the pH and the ionic strength of the solution, changing the length and the aromaticity of the molecular wire, as well as cross-linking cytochrome c6 to the rebuilt PS I promises to increase the rate of hydrogen generation by the PS I/molecular wire/Pt nanoparticle bioconjugate. The best rates of H2 generation to-date are at pH 7.0 in 10 mM MgCl2 and 10 mM NaCl using chemically-cross linked cytochrome c6.

back to top Back to Top


The ability to photocatalytically generate clean hydrogen fuel has the potential to satisfy the global demand for storable energy and has advantages over the costly and environmentally unfriendly current means of hydrogen production. Biologically inspired systems that employ PS I as a means of photon capture and as a source of reducing electrons are yielding promising results when attached to Pt nanoparticles for the light-driven generation of hydrogen.

back to top Back to Top




  1. Rosen, M.A., Scott, D.S., Int. J. Hydrogen Energy, 1998, 23, 653.
  2. United Nations Development Program, World Energy Assesment Report: Energy and the Challenge of Sustainablity, 2003, United Nations, New York.
  3. Mizukoshi, Y., Makise, Y., Shuto, T., Hu, J., Tominaga, A., Shironita, S., Tanabe, S., Ultrasonics Sonochem., 2007, 14, 387.
  4. Gurunathan, K., Maruthamuthu, P., Sastri, M.V.C., Int. J. Hydrogen Energy, 1997, 22, 57.
  5. Dhanalakshmi, K.B., Latha, S., Anandan, S., Maruthamuthu, P., Int. J. Hydrogen Energy, 2001, 26, 669.
  6. Tang, H., Berger, H., Schmid, P.E., Levy, F., Solid State Commun., 1994, 92, 267.
  7. Melis, A., Zhang, I., Forestier, M., Ghirardi, M.L., Seibert, M., Plant Physiol., 2000, 122, 515.
  8. Schutz, K., Happe, T., Troshina, O., Lindblad, P., Leitao, E., Oliveira, P., Tamagnini, P., Planta, 2004, 218, 350.
  9. Ihara, M., Nishihara, H., Yoon, K.S., Lenz, O., Friedrich, B., Nakamoto, H., Kojima, K., Honma, D., Kamachi, T., Okura, I., Photochem. Photobioil., 2006, 82, 676.
  10. Lee, J.W., Tevault, C.V., Blankinship, S.L., Collins, R.T., Greenbaum, E., Energy & Fuels, 1994, 8, 770.
  11. Millsaps, J.F., Bruce, B.D., Lee, J.W., Greenbaum, E., Photochem. Photobiol., 2001, 73, 630.
  12. Evans, B.R., O’Neill, H.M., Hutchens, S.A., Bruce, B.D., Greenbaum, E., Nano Lett., 2004, 10, 1815.
  13. Brettel, K., Bichim. Biophys. Acta, 1997, 1318, 322.
  14. Jordan, P., Fromme, P., Witt, H.T., Klukas, O., Saenger, W., Krauss, N., Nature, 2001, 411, 909.
  15. Gibbs, M., Holaender, A., Kok, B., Krampitz, L.O., San Pietro, A., A report on a workshop held September 5-6, 1973, at Bethesda, MD, supported by NSF under RANN Grant GI 40253 to Indiana University.
  16. Shinkarev, V.P., Vassiliev, I.R., Golbeck, J.H., Biophys. J., 2000, 78, 363.
  17. Antonkine, M.L., Maes, E.M., Czemuszewiez, R.S., Breitenstein, C., Bill, E., Falzone, C.J., Balasumbramanian, R., Lubner, C.E., Bryant, D.A., Golbeck, J.H., Biochim. Biophys. Acta, 2007, 1767, 712.
  18. Que, L. Jr., Bobrik, M.A., Ibers, J.A., Holm, R.H., J. Am. Chem. Soc., 1974, 96, 4168.
  19. Grimme, R.A., Lubner, C.E., Bryant, D.A., Golbeck, J.H., J. Am. Chem. Soc., 2008, 130, 6308.

back to top Back to Top

Related Links