Scientists are taking a leaf out of Nature's book by attempting to harness enough energy from sunlight to carry out useful photochemical reactions. Philip Ball and David Andrews take up the story.

Scientists are taking a leaf out of Nature’s book by attempting to harness enough energy from sunlight to carry out useful photochemical reactions. Philip Ball and David Andrews take up the story.

The energy industry and environmental campaigners don’t tend to agree on much, but there is one thing they both want: hydrogen. As a fuel, hydrogen contains three times the energy of the equivalent mass of hydrocarbons, and its combustion produces neither greenhouse gases nor toxic byproducts - only water. Moreover, the water that covers three quarters of the planet’s surface is its simplest chemical source.

Unfortunately, stripping hydrogen from its bonds to oxygen in H2O requires energy. Current methods for splitting the two elements - principally electrolysis and steam reforming - generally depend on energy from conventional fossil-fuel sources. The alluring dream of advocates of the ’hydrogen economy’ is a cheap and efficient means of splitting water with sunlight.

Since over a thousand times more solar energy reaches the Earth’s surface than is required to power all current human activities, that goal seems to make a lot of sense. But photovoltaic cells remain too expensive to make large-scale solar electrolysis of water feasible. Nature, on the other hand, uses sunlight to split water abundantly and spontaneously in photosynthesis. As the detailed structures and mechanisms involved in this process become better understood, we can ask whether nature has a thing or two to teach us about how to convert light to chemical energy - perhaps not just for hydrogen synthesis, but for powering all kinds of chemical transformations.

But very little (about 4 per cent) of the sunlight delivered to Earth is in the ultraviolet range - so little that no biological system could afford to be wholly dependent on it. Much more solar energy (43 per cent) arrives in the lower-energy form of visible light; it is this that plants and bacteria capture to power photosynthesis. Somehow they manage to split water with light, even though the individual light packets (photons) would seem to be too feeble for the job.

It’s no mystery how photosynthetic organisms manage it: they use more than one photon. In fact, it requires four photons to split one molecule of water into hydrogen and oxygen during photosynthesis. In general, organisms release the oxygen but save up the hydrogen atoms as ions, creating a ’proton gradient’ across the membrane in which the photosynthetic apparatus is anchored. This electrochemical gradient then drives the enzyme ATP synthase, which makes energy-rich ATP molecules - power packs that are tapped for the energy needed to make sugars from carbon dioxide.

The trick, then, is to pool the visible-light photons; to save them up until collectively they have enough energy to do something that none can do alone. Photosynthesis, both in its natural form and in biomimetic chemical systems, is not simply about capturing photons; it is about harvesting light, gathering the solar energy by channelling it from photon-absorbing molecular groups (chromophores) into a reservoir. Such light harvesting also has the advantage that it allows photons to be gathered over a wide area; the photons whose energy is to be combined do not have to arrive at precisely the same location.

Some of these accessory pigments absorb light at wavelengths where chlorophyll absorbs rather poorly - in the blue-green, yellow and orange parts of the spectrum. The photosystem’s full antenna array therefore provides good spectral as well as spatial coverage.

For the array to work properly, all pigments must channel their absorbed energy towards the reaction centre. The energy can pass between molecules thanks to a process called resonant energy transfer. If the molecules are close enough together - typically within a few nanometres - then they are coupled like two pendulums attached to the same beam: if one starts oscillating, it can set the other moving. This coupling depends not only on distance but also on the relative orientation of the two chromophores. Readers who recall the days of moving set-top television aerials around to get a decent picture will understand the principle exactly.

At a fundamental level, we can think of resonant energy transfer in terms of one chromophore emitting a photon that another chromophore absorbs. In fact, work by one of us (David Andrews) and colleagues at the University of East Anglia, UK, has shown that the proper quantum mechanical description of the process involves a ’virtual photon’ passing from one chromophore to the other. Since some energy is lost as heat in the short interval between successive transfer processes, each step ’downgrades’ the solar energy a little, shifting it to longer wavelengths. This means that there is directionality to the energy-transfer process: the energy flows ’downhill’, towards chromophores that absorb at progressively longer wavelengths.

In the late 1990s, Raoul Kopelman and his colleagues at the University of Michigan, US, decided that, to make artificial molecular systems that harvest light, not only could they learn from photosynthesis - they could do better. The antenna array and the use of a ’spectroscopic gradient’ for funnelling energy are marvellous ideas. But the antenna pigment molecules of a plant’s photosystem are packed together in a rather disorderly manner, so that the path the energy must take to reach the reaction centre is to some extent random.

How much better it would be, Kopelman thought, if the chromophores were ordered so that the gradient was shaped like a true funnel, channelling energy steadily from the periphery to the centre of the array. In fact some photobacterial systems operate in just this fashion.

Rather than trying to arrange individual light-absorbing molecules into such a configuration, Kopelman’s group realised they could combine all the chromophores in a single giant molecule: a dendrimer. These nanoscale macromolecules are highly branched polymers with precisely controlled architectures. They are made from branching monomeric units that spread out hierarchically from a central focus, earning them the alternative soubriquet ’nanostars’.

Kopelman and colleagues realised that they could create a covalently bound antenna array by making dendrimers from monomers that contain suitable chromophores. And they could give it a spectroscopic gradient by using monomers in successive generations - each ’layer’ of the macromolecule - with progressively blue-shifted absorption peaks. That is to say, the second-generation chromophores need to absorb at slightly shorter wavelengths than those in the first generation, and so on. Then, photons absorbed by the outermost layer would be passed by resonant energy transfer down the gradient to the centre of the molecule.

The basic structural unit of the monomers the researchers used was phenylacetylene: a benzene ring with an ethyne group attached. These units can be linked together into linear, relatively rigid rods. The conjugated electron clouds in the rods have excitation energies in the ultraviolet/visible part of the spectrum. By adding progressively more phenylacetylene units to the arms of a monomer, the absorption maximum shifts to ever lower energies. The first generation monomers that Kopelman and colleagues used in their dendrimers were rods with four phenylacetylene units: three of these rods were attached to a central benzene ring. The second generation rods (six of them) had three phenylacetylenes, then two, then one in the outermost tips of the dendrimer.

The researchers first showed in 1997 that these molecules did indeed act as energy funnels. When a fluorescent ’reporter’ group was attached at the central focus, ultraviolet light absorbed by the outer chromophores made the reporter group ’light up’ with longer-wavelength blue light. The efficiency with which the energy was delivered to the reporter group was a hundred times greater in the funnelled dendrimers than in versions in which all the chromophore monomers were identical, with no energy gradient.

Normally this molecular unit would be quite unaffected by infrared photons, which carry far too little energy to induce such photoisomerisation. But in the dendrimer, energy was pooled and channelled (at random - there was no energy gradient) into the core, where it caused the azobenzene to switch its shape.

Aida now heads a five-year government-funded scheme in Japan called the Nanospace project, which aims to explore photochemical reactions within light-sensitive dendrimers. Because the density of branching increases with each generation of monomer units, the inside of a dendrimer can be relatively empty, creating a cavity of free space that is protected by the dense thicket of outer branches.

Small molecules can be encapsulated in this free space, and Aida’s results suggest that light energy absorbed in the branches can build up within the dendrimer, shielded from any molecular interactions that would dissipate it as heat, until there is enough energy to initiate a photochemical transformation of the entrapped molecule. The Nanospace project aims to develop biomimetic light-harvesting systems that generate hydrogen.

Recently the research groups of Fritz V?gtle at the University of Bonn, Germany, and Vincenzo Balzani at the University of Bologna, Italy, showed that the energy gathered by a dendrimer’s branches can indeed be channelled into encapsulated molecules. They built a dendrimer that had 64 chromophores of three different kinds. The wavelength of the light absorbed within it got longer, changing from ultraviolet to blue-green as it passed from the outer to the inner groups of the dendrimer, in turn setting up a spectroscopic gradient. A molecule of the red dye eosin (which absorbs green light) was trapped in the core of the dendrimer, and when this system was illuminated with ultraviolet light, the absorbed energy was channelled into the eosin with at least 80 per cent efficiency.

Ken-ichi Sugiura and coworkers at Osaka University in Japan have created square arrays almost 7 nm across, each consisting of 21 porphyrins linked in a kind of four-armed snowflake structure. And recently, another Japanese group led by Atsuhiro Osuka at Kyoto University, succeeded in making linear arrays of zinc porphyrins up to 24 units long, ending with an acceptor group that receives and re-radiates most (sometimes all) of the absorbed energy.

Jonathan Lindsey at North Carolina State University in Raleigh, US, has been exploring supramolecular systems like this for several years. In 1999 he and his coworkers teamed up with artificial-photosynthesis enthusiasts Ana and Thomas Moore and Devens Gust at Arizona State University in Tempe, US, to make an antenna complex in which four zinc porphyrin groups covalently linked in a T shape were attached to a ’free-base’ (FB) porphyrin - which has no metal ion in the middle. This fifth porphyrin was in turn hooked up to a C 60 molecule . The idea was that the zinc porphyrins would act as the antenna array, capturing light energy and channelling it into the unit containing both the FB porphyrin and the C60, which serves as a reaction centre.

The researchers indeed found that when the zinc porphyrins were selectively excited by yellow light, the energy was transferred within 300 picoseconds (300 x 10-12 s) onto the FB porphyrin, where it caused an electron to hop to the C 60 molecule. This turns the two molecular groups in the reaction centre into charged free radicals, which are potentially very reactive and, as such, useful initiators for further chemistry.

Last year, Lindsey and the Arizona State team showed that by tinkering with the chemical structure of the FB porphyrin they could make the radical ions about 200 times longer-lived - they survive for 240 nanoseconds before the electron hops back. The charge-separated state was also formed more efficiently, with 90 per cent probability each time one of the zinc porphyrins absorbs a photon.

The big challenge now is to persuade these light-harvesting arrays to spark off some useful chemistry. It is already clear that energy can be gathered and channelled to specific locations, and even that it can be pooled so that low-energy photons can be put to useful work. But can we really split water this way, and generate the green fuel of the future economy? With the current rate of progress, that looks like a fair bet.


Philip Ball is consultant editor for Nature, 4-6 Crinan Street, London N1 9XW; and David Andrews is professor of chemistry in the school of chemical sciences and pharmacy, University of East Anglia, Norwich NR4 7TJ.

Bugs make light work

The photosystem of green plants isn’t the only natural model for artificial light-harvesting systems. A number of specialised bacteria are capable of photosynthesis too, but they do it rather differently. In the anaerobic depths of certain lakes where many of them live, they need to be far more efficient than plants, because they must harvest energy from very little light. The largest phylum of bacteria is called Proteobacteria, or purple bacteria, all of which are thought to have stemmed from an ancestral photosynthetic species.

Purple photosynthetic bacteria have two light-harvesting systems, LH1 and LH2, both of which are circular complexes of proteins and pigment molecules. The important water-splitting reactions begin in a reaction centre in LH1. Each LH1 complex is surrounded by several (smaller) LH2 complexes, which act as an antenna system that collects and channels light energy. The porphyrin chromophores in these complexes are housed on a type of chlorophyll pigment called bacteriochlorophyll.

In both LH1 and LH2 the porphyrin units are stacked around the circumference of the ring-like assemblies. If any of the 32 porphyrins in the LH2 complex absorbs a photon, the energy can be shunted between neighbouring porphyrins, creating a kind of energy storage ring. It can be tapped at any point around this ring and transferred to the LH1 complex.

This elegant supramolecular architecture has inspired several synthetic mimics. In 1999, Jonathan Lindsey and coworkers at North Carolina State University in Raleigh, US, made a barrel-shaped ring of six linked porphyrins: three of them free-base versions and three containing zinc. They assembled porphyrin building blocks around a central molecular template - a procedure pioneered by Jeremy Sanders at the University of Cambridge, UK. The final ring measures about 1.9 nm across.

The researchers found that energy could be transmitted very quickly and efficiently from the zinc to the FB porphyrins, mimicking the process in the LH2 complex itself. Recently, Ryoichi Takahashi at the Nara Institute of Science and Technology and Yoshiaki Kokube at the Japan Science and Technology Corporation built a still bigger ring, about 4 nm in diameter, from six zinc-porphyrin dimers, making a total of 12 porphyrins in all.

In this array, adjacent porphyrins overlap with one another rather as they do in the LH2 ring, which might make energy transfer more efficient, although the researchers have yet to investigate how light energy gets absorbed and shunted around.