Controlling the Work Function of Indium Tin Oxide: Differentiating Dipolar from Local Surface Effects Eric L. Bruner, Norbert Koch, Amelia R. Span, Steven L. Bernasek, Antoine Kahn, ‡ and Jeffrey Schwartz* Department of Chemistry, Princeton UniVersity, Princeton, New Jersey 08544-1009, and Department of Electrical Engineering, Princeton UniVersity, Princeton, New Jersey 08544-5263 Received October 9, 2001 Because of the ubiquitous use of indium tin oxide (ITO) as an anode material for optoelectronic devices, considerable effort has gone into the development and testing of models aimed at enhancing hole injection from ITO into an organic overlayer. Among these models 1 is an elegant one based on simple concepts of electrostat- ics: 2 organization of a dipole layer at the surface of an ITO electrode is predicted to affect its work function (φ) linearly, and ultimately its carrier injection ability. 1 This model predicts that organizing a dipolar layer on an electrode surface with its negative end furthest from that surface should increase φ of that electrode. 2,3 Impressive tests of the dipole model have been conducted with use of self- assembled monolayers of para-substituted arylthiols on gold 3 or copper 4 and para-substituted benzoic acids on CdTe, 5 CdSe, 6 GaAs, 7 TiO 2 , 8 and ITO. 9 However, such tests suffer from a degree of ambiguity in that para-substitution in these surface modification reagents affects not just their molecular dipole moments, but the electronegativity of their ligating groups (thiolate or carboxylate) analogously. Thus, attempts to understand, and ultimately exploit, para-substituted aryl reagents for control of φ are compromised: Local electronegativity effects, including proton transfer from the reagent to the electrode surface, cannot be excluded as causative for any changes in φ observed. (Also problematic is that any proton transfer from an acid to an electrode surface could have a deleterious effect on device performance. 10,11 ) We describe herein surface modification of ITO using organotin complexes which enables systematic control of the ITO electrode work function and which obviates the problem of net proton transfer to the ITO surface. 12,13 Furthermore, our system enables clear differentiation between the effects of group dipole and electronegativity of our surface modification reagents on the measured change in φ for ITO. We have described 12 the reaction between surface OH groups of ITO and tetra(tert-butoxy)tin (1), which gives surface tin alkoxides. Tin alkoxide 1 (Strem; 90L exposure) was deposited on the hydroxylated ITO, which was held at 150 K; the sample was then warmed to 293 K to give [ITO]-[O] 2 -Sn(OBu t ) 2 /[ITO]-[O] 3 - Sn(OBu t )(2/3), as determined by X-ray photoelectron spectroscopy (XPS) from surface C:Sn ratios (see Table 1). 12,13 Next, a series of substituted phenols was vapor deposited onto 2/3 (approximately 90 L, with the substrate held at 150 K). Substrates were then heated to 290 K (except for p-nitrophenol, which was heated to 390 K) to enable ligand metathesis and to desorb multilayer phenol. 13 Sto- ichiometries of the tin phenoxide surface complexes formed, [ITO]- [O] 2 -Sn(OC 6 H 4 -X) 2 /[ITO]-[O] 3 -Sn(OC 6 H 4 -X) (4/5; a,R 1 ) R 2 ) R 3 ) H; b,R 1 ) R 3 ) H, R 2 ) OCH 3 ; c,R 1 ) R 3 ) H, R 2 ) F; d,R 1 ) R 3 ) H, R 2 ) CF 3 ; e,R 1 ) R 3 ) H, R 2 ) NO 2 ; f,R 1 ) R 3 ) F, R 2 ) H; g,R 1 ) OCH 3 ,R 2 ) R 3 ) H; h,R 1 ) F, R 2 ) R 3 ) H), were determined by XPS (Table 1; Scheme 1); no appreciable change in ratios of starting materials 2:3 to products 4:5 was measured, except for 4e/5e, because of heating. 12,13 XPS analysis of F, N, organic O, or fluorinated alkyl C provided a complementary determinant of surface complex stoichiometries. Using the In(3d 5/2 ) peak as an internal standard allowed relative changes in surface tin complex loading to be quantified. Measure- ment of the In:Sn ratio for each sequence of deposition of 1 and ligand exchange determined that these complex loadings were consistent among experiments within 25%. Changes in φ for ITO on substitution were also measured by XPS. Secondary electrons in photoemission are produced by inelastic scattering of primary electrons within the sample. Electrons with a kinetic energy equal to or larger than the vacuum level of the sample can be detected. A negative bias was applied to the sample, ensuring that all low kinetic energy electrons (including those possessing “zero” kinetic energy) could be detected. The work function is obtained from eq 1, in which the incident photon energy (hν), the kinetic energy of electrons emitted from occupied states at the Fermi energy (E EF ), and the low kinetic energy onset of the secondary electron distribution (E onset ) are related. 1 E EF is constant when the photon energy is fixed and the sample and spectrometer are in thermodynamic equilibrium, in electrical contact; therefore, a change in the sample φ can be determined by measuring the corresponding shift in the low kinetic energy onset of the secondary electrons. Resolution in this experiment is determined by the pass * Address correspondence to this author at the Department of Chemistry, Princeton University. E-mail: jschwartz@chemvax.princeton.edu. ‡ Department of Electrical Engineering. Table 1. Gas-Phase Dipole Moments and Measured Changes in φ for 4/5 phenol μz phe a Δφ b phenol μz phe a Δφ b a -0.44 0.47 e 5.14 1.11; 1.08 b -0.55 0.48 f -1.85 0.29 c 1.20 0.67 g -1.47 0.30 d 3.06 0.92; 0.87 h -1.18 0.29 a Dipole moments 15 are reported along the z-axis, defined as the C-O bond of the phenol. b Work function changes are between measured values of φ for the tin phenoxides (4/5) versus the alkoxides (2/3) for each sample. Scheme 1. Synthesis of ITO Surface Bound Tin Phenoxides Published on Web 03/06/2002 3192 9 J. AM. CHEM. SOC. 2002, 124, 3192-3193 10.1021/ja012316s CCC: $22.00 © 2002 American Chemical Society