a strong decay channel in our device. For triplet±triplet annihilation to be strong, we require electron±hole recombination to be confined to a very small region inside the polymer layer, where a sufficiently high triplet density can build up. This situation is certainly possible in our hole-injection limited F8BT LED. [18] However, we would expect such triplet interactions to produce long-lived singlet emission with a lifetime comparable to the lifetime of the triplet. No such emission was observed in our experiments. We therefore consider the bimolecular decay of triplets to be a relatively weak effect in our devices. In principle (iii), a small triplet cross-section value, could explain our data. The value, however, would have to be at least two orders of magnitude smaller than cross-sections found in other polymers to be consistent with singlet±triplet ratios found in other devices, which is highly unlikely. For this reason, we conclude that (i), a very small triplet formation probability, is the most likely explanation for the absence of a triplet absorption in our F8BT devices.To understand why triplet formation is strongly suppressed in F8BT compared to F8 and PPV, one can focus on the differing physical properties of singlet and triplet excitons arising from electron-correlation effects. As a carrier hops onto a chain harboring an oppositely charged carrier, the charges can recombine or migrate away from each other. As discussed recently, the exchange interaction between oppositely charged carriers is governed by the extent of the conjugation along the polymer backbone, and hence the chain rigidity. [5,6] The coulombic field is, however, mostly unaffected by the chain conformation. Consequently, spin-symmetric charges residing on opposite ends of a rigid polymer chain will have a larger exchange energy compared to a less rigid chain, and will thus favor formation of singlet excitons. We therefore speculate that the highly planar conformation expected of F8BT compared to F8 and PPV is responsible for the high singlet fraction found in the device. Hence, we propose that improvements in the planarity of the polymer backbone favor smaller triplet yields due to improved mediation of the exchange interaction.Absorptions due to triplet excitons generated in polyfluorene-based polymer LEDs have been measured and limits set on their formation efficiency. Although triplets were found in F8 LEDs, no evidence for triplets was found in F8BT homopolymer or blend devices. This null measurement suggests that exciton formation is not only spin-dependent in conjugated polymers, but is highly spin-dependent in F8BT. Our work therefore highlights dependence of the singlet±triplet ratio on polymer chemical structure, showing that triplet formation can be heavily suppressed and is not a necessary obstacle to high-efficiency device operation. As a consequence, more work needs to be done to assess the ultimate limits of polymer LED efficiency.
ExperimentalThe PA spectroscopy technique used in this work has been described elsewh...