Plasmonic Internal Photoemission for Accurate Device In Situ Measurement of Metal‐Organic Semiconductor Injection Barriers

Abstract

4775www.MaterialsViews.com wileyonlinelibrary.com the challenge of independently and reliably measuring key parameters such as the injection energy barrier directly in the device geometries in which currentvoltage ( J -V ) characteristics are measured.Ultraviolet photoelectron spectroscopy (UPS) is arguably the most successful and widely used method for determining energy level alignment at MO interfaces, [8][9][10] however the energetics deduced from this method as well as its inverse counterpart are not in general the same as that in an actual device due to the difference between surface and bulk polarization energies. [ 9,10 ] That is, the energy of a charge carrier near an MO interface is stabilized by polarization and lattice relaxation of the surrounding molecules and the adjacent metal (≈0.5-1 eV), whereas surface-specifi c techniques such as UPS, Kelvin probe and scanning tunneling microscopy all measure the energy of charge carriers at an organic fi lm surface, stabilized only by polarization from the underlying half-space.Internal photoemission (IPE) by contrast is carried out directly in the device of interest (i.e., device in situ) and is well established for measuring the injection barrier at metalinorganic semiconductor contacts. [ 11,12 ] In this approach, the metal contact is illuminated directly with visible/near infrared light below the semiconductor bandgap leading to the excitation of hot carriers in the metal. Those carriers with suffi cient energy and momenta to surmount the injection barrier, B φ , are emitted into the semiconductor resulting in a photocurrent with yield per absorbed photon, Y , according to the Fowler relationship: [ 13 ] where A is a constant that refl ects the strength of the involved optical transitions. Internal photoemission has previously been applied to MO interfaces, [ 14,15 ] however as demonstrated below, interpretation of the data is complicated by the fact that disorder and impurities in organic semiconductors lead to low-level absorption tails extending deep into the optical gap. This sub-gap absorption, together with photo-detrapping transitions within the disorder-broadened highest occupied (HOMO) and lowest unoccupied molecular orbital (LUMO) transport levels leads to a background photocurrent that is diffi cult to Current injection in organic semiconductors remains diffi cult to predict due in large part to the challenge of characterizing the contact energy barrier and interface density of states directly in organic electronic devices. Here, resonant coupling to surface plasmon polariton modes of a metal contact is demonstrated as a means to carry out internal photoemission (IPE) accurately in disordered organic semiconductor devices and enable direct measurement of the contact injection barrier by isolating true IPE from spurious sub-gap organic photoconductivity. The substantial increase in sensitivity afforded by resonant coupling enables measurement in the low-fi eld injection regime where deviation from the standard Fowler prediction is explained q...

Electrical contacts to two-dimensional semiconductors

“…Other methods that have been used to extract the SBH in 2D SCs include ultraviolet and Xray photoelectron spectroscopy [66][67][68][69] , which measures the band bending, and scan ning tunnelling spectroscopy of Au nanoparticles deposited on top of MoS 2 (refs 70,71). Finally, some techniques that could in future be employed for TMDCs include internal photoemission 72,73 and C-V measurements 30 . Figure 5 plots SBH values between multilayer MoS 2 and different metals 24,60,70,74 against the corresponding metal work functions.…”

Large bipolaron density at organic semiconductor/electrode interfaces

“…The main conclusion from Fig. 5b is that, in the Fermi level pinning regime, bipolarons can generally be expected to make up a substantial fraction of the interfacial charge density ( P 2+ / P + > 0.01) over the entire range of σ i that have been estimated to date 4 , 26 – 28 . This fraction is even higher for alternative interface DOS distributions, such as Lorentzian or exponential functions that decay more slowly than the Gaussian employed here 8 , 28 .…”

“…1 a 12 . We take σ b = 0.1 eV based on bulk transport measurements of the closely related molecule NPD 24 , 25 and σ i = 0.35 eV in accord with estimates of the interfacial distribution obtained from internal photoemission and impedance spectroscopy 26 , 27 .…”