LASERS AND OPTICS Semiconducting polymers are moving out of the research lab and into the market-place as industry realizes the commercial potential of polymer-based LEDs, displays and photovoltaics Polymer diodes Richard Friend, Jeremy Burroughes and Tatsuya Shimoda FROM CAR bumpers to bullet-proof Bonding basics vests – everyone is familiar with poly- The common feature of all carbon- mers as flexible yet mechanically based molecular semiconductors is strong materials. Less well known is that they consist of alternate single the fact that some polymers can also and double carbon–carbon bonds. conduct electricity and emit light. Such materials are said to be “conju- These semiconducting polymers, gated” if it is possible to swap the posi- which have been intriguing research- tions of the single and double bonds ers for the past 20 years, are now and end up with a structure that still poised to enter the market-place. One satisfies the chemical-bonding re- of the most advanced applications quirements for carbon. Carbon– lies in displays, where semiconducting carbon double bonds are formed polymers can be used as the active when two of the three 2p orbitals on element in light-emitting diodes each carbon atom combine with the (LEDs). Philips, the Dutch electronics 2s orbital to form three sp2 “hybrid” giant, has moved fastest to the market, orbitals. These orbitals lie in a plane, and will commission a polymer-LED directed at 120o to one another, and pilot production line at its factory in form three s “molecular” orbitals Heerlen in the Netherlands later this with neighbouring atoms, including year. The firm will probably use the one with carbon. The third p orbital devices as back-lights for mobile on the carbon atom, pz, points per- phones and other portable electronic pendicular to the other hybrid or- products. Another promising appli- bitals. It overlaps with the pz orbital on cation of semiconducting polymers is neighbouring carbon atoms to form a in the field of photovoltaics, where Plastic fantastic – a polymer LED display in which the pair of so-called p molecular orbitals they could be used as area detectors pixels have poly(phenylene vinylene) as the active element that are spread out or “delocalized” or solar cells. over the polymer chain. The lower- The main reason for all this interest is that semiconducting energy p (or bonding) orbital is the valence band, while polymers are easy to process. Semiconductor devices consist the higher-energy p* (or antibonding) orbital forms the of a series of layers – electrodes, the active semiconductor conduction band. This “band gap” gives the polymer its material, insulators and so on – and it is attractive if they can semiconducting behaviour. be assembled, as polymers can, in solution. But a polymer must also satisfy two other conditions for it to Another advantage of polymer-based semiconducting de- work as a semiconductor. One is that the s bonds should be vices is that it is easy to engineer the interfaces between the much stronger than the p bonds so that they can hold the various layers to make working devices. In contrast, it is vital molecule intact even when there are excited states – such as for the surfaces of inorganic semiconductors to be struc- electrons and holes – in the p bonds. (These semiconductor turally regular at an atomic level. Any ruptured chemical excitations weaken the p bonds and the molecule would split bonds at the surface of these materials will have “non-bond- apart were it not for the s bonds.) The other requirement is ing” orbitals that can stop the device from working. However, that p orbitals on neighbouring polymer molecules should this problem does not occur in semiconducting polymers, overlap with each other so that electrons and holes can move which only have covalent bonds within each molecule. This is in three dimensions between molecules. Fortunately many an important advantage, because it allows almost any combi- polymers satisfy these three requirements. nation of semiconducting-polymer materials to be used to Most conjugated polymers have semiconductor band gaps build a device layer by layer. Semiconducting polymers are of 1.5–3 eV, which means that they are ideal for optoelec- also interesting from a fundamental point of view because tronic devices that emit visible light. They can also be chem- their electronic properties lie at the crossroads between tradi- ically modified in a variety of ways, and a lot of effort has been tional inorganic materials, such as silicon, and, for example, put into finding materials that can be processed easily from photoactive biomolecular materials. solution – either as directly soluble polymers, or as “precursor” P HYSICS W ORLD J UNE 1999 35 LASERS AND OPTICS polymers that are first processed in 1 Polymer LEDs energy barrier of 0.4 eV to enter the solution and then converted in situ to polymer layer. form the semiconducting structure. O O However, several groups have The most widely used polymer is found that including a layer of 2.8 eV poly(phenylene vinylene) or “PPV”, S n 2.7 eV conducting polymer between the which has a band gap of about PEDOT indium–tin oxide and the light-emit- 2.5 eV and emits yellow–green light. n ting polymer helps to improve the PPV is made up of benzene rings – injection of holes. (The extra layer n PSS the “phenylene” units – linked to one PPV is deliberately kept thin so that the SO H another via two carbons, which are 3 device remains transparent.) One 4.8 eV each also bonded to a single hydro- 5.0 eV good example is poly(ethylene- gen atom, or “vinylene” unit (figure 5.2 eV dioxy)thiophene, or “PEDOT”, a 1). Researchers often modify PPV by material that was developed by attaching alkyls side-chains, such as Bayer, the German chemical com- indium–tin- semiconducting calcium cathode (CH2)nCH3, to the phenylene rings. oxide anode polymer pany. PEDOT can be mixed with This makes the polymer soluble in poly(styrene sulphonic acid), or convenient solvents, such as xylene. 40 nm 100 nm “PSS”, to form a stable and highly It can then be easily processed A polymer light-emitting diode made with poly(phenylene conducting charge-transfer com- directly from solution, using tech- vinylene), or “PPV”, sandwiched between a calcium cathode plex, and this can be deposited from and an anode made from a conducting polymer on niques such as spin-coating, whereby indium–tin oxide (ITO). The calcium injects electrons into an aqueous solution to form a layer a drop of polymer solution is placed the polymer film, while the anode injects holes. When an on the indium–tin oxide (figure 1). on a rapidly rotating substrate to electron and hole capture one another within the PPV, they The PEDOT/PSS layer has a high form a thin, uniform film. form neutral “excitons” – bound excited states that decay by work function (5.0 eV), which allows emitting a photon of light. The ITO layer is generally formed on a glass substrate, and the device is made by depositing holes to be injected easily into the Polymer light-emitting diodes the various layers in turn onto this substrate. The PPV since the barrier is now much Efficient polymer light-emitting conducting polymer shown here is a derivative of smaller. Another advantage of the diodes (LEDs) can be made from poly(ethylenedioxy)thiophene (PEDOT) doped with PEDOT/PSS layer is that it poly(styrene sulphonic acid) (PSS). very simple structures. The first smoothes out the relatively rough results, which two of the authors surface of the indium–tin oxide, (RF and JB) reported in 1990, were made using a thin layer thereby preventing any local short-circuiting that would of polymer sandwiched between a pair of electrodes. The otherwise cause the device to fail. An added bonus is that the negative electrode injects electrons into the polymer film, surface of the PPV in contact with the PEDOT/PSS is lightly while the positive electrode injects holes. When an electron p-type doped by the PSS that builds up at the interface, which and hole capture one another within the polymer, they form makes hole injection even easier. a neutral “exciton”, which is a bound excited state that can What about the electron-injecting electrode? This should decay by emitting a photon. This process is known as ideally be a metal with a low work function that matches the “electroluminescence” since the emission of light is caused “electron affinity” – the difference in energy between the by the electric field between the electrodes. The negative bottom of the conduction band and the vacuum level – of the electrode is chosen to have a low work function so that elec- PPV.Calcium has been widely used for this purpose. Its work trons can be easily injected, while the positive electrode has a function of 2.8 eV is close to the electron affinity of PPV high work function so that it can “suck in” electrons from the (2.7 eV) and the barrier for electron injection is therefore polymer. The hole-injecting electrode is usually made from small. As with the anode, the chemistry of the interface with indium–tin oxide, which is optically transparent and there- the semiconducting-polymer layer is important. Calcium fore allows the light to leave the device. forms an ionic charge-transfer complex with the surface These early devices were not particularly good. They emit- layers of the PPV,which n-dopes the polymer. Bill Salaneck ted very few photons relative to the number of charges and colleagues from Linköping University in Sweden have injected, i.e.