Ž . Synthetic Metals 111–112 2000 311–314 www.elsevier.comrlocatersynmet Kelvin probe and ultraviolet photoemission measurements of indium tin oxide work function: a comparison J.S. Kim a , B. Lagel b , E. Moons a,1 , N. Johansson c , I.D. Baikie b , W.R. Salaneck c , ¨ R.H. Friend a , F. Cacialli a, ) a CaÕendish Laboratory, Department of Physics, UniÕersity of Cambridge, Madingley Road, Cambridge CB3 0HE, UK b Department of Physics, Robert Gordon UniÕersity, Aberdeen, UK c Department of Physics, UniÕersity of Linkoping, SE-518 83 Linkoping, Sweden ¨ ¨ Abstract Ž . We report a comparison of the work functions of thin films of indium tin oxide ITO , carried out by means of ultraviolet Ž . photoelectron spectroscopy UPS and by measurements of the contact potential difference with respect to a gold reference electrode Ž Ž . . Kelvin probe KP method . We investigated commercially available ITOs both ‘‘as-received’’, and after certain surface treatments, such as oxygen plasma. First, we find measurable discrepancies between KP values measured with three different instruments, and between the KP and the UPS values. Secondly, and unexpectedly, we find that the KP, although more sensitive than UPS, does not detect certain differences between ITOs with different surface treatments. We discuss the results in view of the different environments in which the Ž . measurements are carried out UHV for the UPS and airrAr for the Kelvin method , of the effects which may be induced by the high-energy photon irradiation in the UPS measurement, and of the stability of the gold probe work function in gas ambient. We conclude that UPS is better-suited for absolute work function determination, although KP remains a convenient and inexpensive tool for fast screening of contact potential differences. q 2000 Elsevier Science S.A. All rights reserved. 1. Introduction Ž . Indium tin oxide ITO is the common choice as the Ž . anode material of organic light-emitting diodes LEDs , due to its good transparency, low resistivity, and ease of patterning. Among other ITO parameters, the work func- tion value plays a central role in determining efficiency and performance of the LEDs, via control of the hole injection process, as we have shown recently for differ- Ž . ently treated ITO surfaces in poly p-phenylene vinylene Ž . wx PPV -based LEDS 1 . A low ITO work function results in a relatively high barrier for injection of holes into the Ž . highest occupied molecular orbitals HOMO of organic, electroluminescent semiconductors, which lie at 5.0 eV or more, below vacuum. wx As is well-known 2 , however, the work function is an extremely sensitive indicator of the state of a surface, and ) Corresponding author. Ž . E-mail address: fc10004@cam.ac.uk F. Cacialli . 1 Present address: Cambridge Display Technology, Greenwich House, Madingley Road, Madingley Rise, Cambridge CB3 0HJ, UK. therefore of the environment surrounding it, since gases or other adsorbed species may induce substantial variations of the energy required to remove an electron from the Fermi level. This is true, in particular, for metal oxides. In addition, the measured value is also known to be a func- tion of the technique used. The Kelvin method measures wx the average work function difference under the probe 3, Ž . whereas ultraviolet photoelectron spectroscopy UPS mea- sures the lowest work function patch on the surface, even if it is only a small fraction of the surface area. Thus, Ž . Kelvin probe KP work function values are often higher than those measured by UPS. These factors have important consequences on the way various measurements and published data are used for designing organic LEDs. In particular, they affect the use of specific combinations of electroluminescent semicon- ductors and electrodes, chosen in order to reduce to a minimum the energy barriers for electronrhole injection into the semiconductor. Here we report a comparison between measurements conducted in an ultra-clean environment under irradiation Ž . with 21.22 eV photons UPS with dark measurements in 0379-6779r00r$ - see front matter q 2000 Elsevier Science S.A. All rights reserved. Ž . PII: S0379-6779 99 00354-9