The Impact of Local Work Function Variations on Fermi Level Pinning

Article pubs.acs.org/JPCC

The Impact of Local Work Function Variations on Fermi Level Pinning of Organic Semiconductors Stefanie Winkler,†,‡ Johannes Frisch,‡ Raphael Schlesinger,‡ Martin Oehzelt,†,‡ Ralph Rieger,§ Joachim Rad̈ er,§ Jürgen P. Rabe,‡ Klaus Müllen,§ and Norbert Koch*,†,‡ †

Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Albert-Einstein-Straße 15, 12489 Berlin, Germany Humboldt-Universität zu Berlin, Institut für Physik, 12489 Berlin, Germany § Max Planck Institute für Polymerforschung, Ackermannweg 10, 55128 Mainz, Germany ‡

S Supporting Information *

ABSTRACT: This photoemission study shows that the work function (Φ) of indium−tin-oxide (ITO) can be increased from 4.2 up to 6.5 eV upon the deposition of the molecular electron acceptors tetrafluoro-tetracyanoquinodimethane (F4TCNQ) and hexaazatriphenylene-hexacarbonitrile (HATCN). The evolution of sample Φ and the hole injection barrier upon subsequent deposition of the hole transport material N,N′-bis(1-naphthyl)-N,N′-diphenyl-1,1′-biphenyl-4,4′-diamine (α-NPD) was studied for different acceptor precoverages of ITO, corresponding to different initial Φ values. When Φ of the acceptor covered substrate exceeds a critical value Φcrit, the highest occupied molecular level of multilayer α-NPD is found to be pinned 0.5 eV below the Fermi level (EF). Noteworthy, Φcrit is found at 5.2 eV, which is 0.4 eV higher than expected for α-NPD (4.8 eV), and vacuum level alignment does not apply even before EF-pinning sets in. An electrostatic model that accounts for nonuniformity of the substrate at acceptor submonolayer coverages and the associated local work function changes explains the origin of “delayed” EF-pinning.

■

INTRODUCTION For efficient injection of charges in organic (opto-) electronic devices, proper matching of the electrode Fermi level (EF) to the charge transport levels of the organic semiconductor is necessary.1−3 Therefore, electrode materials are chosen in a way to minimize that mismatch. Assuming no electronic coupling at the material interfaces, one would expect vacuum level alignment upon contact formation, also known as the Schottky−Mott limit. In this case, it is possible to predict the energy levels at the interface simply from knowing the separate values of the electrode work function (Φ) and the ionization energy (IE) and electron affinity (EA) of the organic semiconductor. For organic/metal interfaces, the validity of vacuum level alignment has been ruled out by the systematic observation of interface dipoles.1,4 In the simplest case, molecules are physisorbed on a metal, where they push back electrons tailing out of the metal due to Pauli repulsion5−8 referred to as “pushback effect”. More complex interactions can occur as well, and generally charge density rearrangements occur at the interface, e.g., by chemical bond formation, charge transfer, and molecular distortions. For most conductive polymers and ambient-contaminated metal and conductive oxide electrodes,9,10 it was found experimentally that vacuum level alignment holds for a certain range of the substrate work function Φsub values, which is specific for every organic semiconductor.3,11 Exceeding the limits of this range, referred to as critical work function Φcrit, vacuum level alignment would place the electrode EF below the © XXXX American Chemical Society

highest occupied molecular orbital (HOMO) level or likewise above the lowest unoccupied molecular orbital (LUMO) level of the organic semiconductor, corresponding to an electronic nonequilibrium situation. In response, charge is transferred, and interface dipoles form such that EF comes to lie within the HOMO−LUMO gap of the organic semiconductor and equilibrium is re-established; this phenomenon is commonly referred to as EF-pinning.3,12 The work function at which EFpinning occurs should in principle be an intrinsic material parameter. However, for one organic semiconductor, often different values for pinning have been reported.13 In fact, research on EF-pinning in organic semiconductors has started only quite recently, and more than one mechanism may be at work to determine the energetics of pinning, and thus more detailed information about the relation between substrate work function and energy level pinning is needed. Within the limiting Φsub value interval, one might predict the energy level alignment, and the “free movement” of EF allows for proper matching of the energy levels at the interface by simply choosing an adequate substrate Φsub. One way to adjust Φsub to an optimal value is to use thin interlayers of strong molecular acceptors14 (donors15), that form tailored interface dipoles to increase/decrease Φsub. In doing so, one can Special Issue: Ron Naaman Festschrift Received: February 24, 2013 Revised: April 26, 2013

A

dx.doi.org/10.1021/jp401919z | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Figure 1. (a) Evolution of Φsub for ITO as a function of F4TCNQ(squares)/HATCN(circles) nominal coverage θacc obtained from UPS. (b) Evolution of the HOMO-onset BE for nominally 5 nm thick α-NPD-films and the sample work function Φα−NPD as a function of Φsub. To adjust Φsub different θacc has been used. EF-pinning occurs at 0.4 eV higher initial work function (Φsub= Φcrit = 5.2 eV) than expected. (c) Molecular structures of the acceptors F4TCNQ (i), HATCN (ii), and the hole transport material α-NPD (iii).

■

EXPERIMENTAL DETAILS Photoemission experiments were performed at an ultrahigh vacuum (UHV) apparatus consisting of interconnected sample preparation (base pressure: 1 × 10−10 mbar) and analysis (base pressure: 1 × 10−10 mbar) chambers. ITO covered glass substrates (sheet resistance 15−30 Ω) were cleaned in situ via repeated cycles of Ar-ion sputtering (500 eV, 1 μA) and annealing at 400 °C (last annealing step performed at 400 °C in 10−6 mbar O2-pressure for 10 min). The absence of sample contamination was confirmed by X-ray photoelectron spectroscopy (XPS) prior to evaporation of the acceptors. Thin film preparation of F4TCNQ/HATCN/α-NPD was performed via evaporation from resistively heated quartz crucibles using a rate of about 0.2−2 Å/min. The film mass thickness of F4TCNQ (density:1.64 g/cm3)/ HATCN(1.6 g/cm3)/ α-NPD (1.23 g/ cm3) was monitored using a quartz crystal microbalance. Ultraviolet photoelectron spectroscopy (UPS) was performed using a helium-gas-discharge lamp (21.218 eV) with very low photon flux (ca. 100 times attenuated compared to standard commercial sources) in order to avoid irradiation damage of the samples. X-ray photoelectron spectroscopy (XPS) was performed using Al/Kα radiation (1486.7 eV). All spectra were recorded at room temperature and normal emission using a hemispherical Specs Phoibos 100 energy analyzer with 120 meV energy resolution for UPS. To determine the work function, secondary electron cutoffs were recorded with the sample biased to −10 V to clear the analyzer work function.

significantly alter the level alignment toward any organic semiconductor on top, and therefore the charge injection properties at the interface. On metals, Φsub increases upon the deposition of an acceptor, such as 2,3,5,6-tetrafluorotetracyanoquinodimethane (F4TCNQ) or hexaazatriphenylene-hexacarbonitrile (HATCN); the structures of both are shown in Figure 1. The origin is a net electron transfer from the metal to the molecule, resulting in an interface dipole. The EA of the acceptors being higher than the work function of the metal provides the driving force for this charge transfer. In analogy, acceptors such as F4TCNQ can also be used to increase the work function of conductive oxides, such as indium−tin-oxide16 (ITO). In this study we chose the amorphous organic semiconductor N,N′-bis(1-naphthyl)-N,N′-diphenyl-1,1′-biphenyl-4,4′-diamine (α-NPD), a widely used hole transport material, for investigating in detail the transition point between vacuum level alignment and EF-pinning at Φcrit. We employed varying acceptor coverages θacc (for both HATCN and F4TCNQ) to increase Φsub incrementally. In the pinning regime, the HOMO-onset of multilayer α-NPD is found to be pinned 0.5 eV below EF with a saturated work function Φα−NPD of 4.8 eV. Assuming vacuum level alignment, pinning is thus expected to start at Φsub of 4.8 eV accordingly, as also reported in other studies.17 By contrast, our results show that EF-pinning set in only at Φcrit = 5.2 eV. At this point, the surface consists of a mixture of organic (acceptor-covered fraction) and inorganic (uncovered fraction) areas, since a full monolayer of the acceptors would cause even higher work function values. With a simple model including this nonuniformity of the substrate, we can explain the “delayed” occurrence of EF-pinning, which is independent of the specific acceptor employed to tune Φsub.

■

RESULTS AND DISCUSSION

The evolution of Φsub versus the nominal acceptor coverage (θacc) is plotted in Figure 1 a). The bare ITO work function is 4.2 eV. When depositing the acceptor molecules, an almost linear increase of Φsub to maximum values of 5.8 eV (HATCN) B



Article

they are not directly distinguishable in spectra19 due to, e.g., small energy splitting or inhomogeneities, one would at least expect significant spectral broadening at low α-NPD film thickness compared to thick films. If, on the other hand, acceptors adsorb on ITO in a “hit-and-stick” fashion, i.e., bare and acceptor-covered ITO areas coexist on a molecular length scale, the second layer of α-NPD already aligns to the averaged work function of the first layer.19,20 Since the spectral shape variations of the α-NPD-HOMO (see Supporting Information) as a function of Φsub (and Θacc likewise) are negligible, we assume local work function variations of acceptor-covered and bare ITO patches to occur on a molecular length scale. In summary, two effects are at play in the studied cases: the work function shift due to α-NPD EF-pinning ΔΦpin (for Φsub>Φcrit) on acceptor-covered areas and the remaining pushback-effect (PB rem ) on bare ITO, whose relative contributions to the overall work function shift depend on Θacc. PBrem, caused by the α-NPD molecules adsorbing on ITO areas not covered by acceptor molecules should thus be the origin of the delayed set-in of EF-pinning, as observed in Figure 1b. To test this proposition, a simple model is applied, which is based on electrostatics and assumes electronic equilibrium between the different layers.18,21 To account for PBrem, we assume a linear correlation between acceptor coverage and pushback (as justified by the almost linear dependence of Φsub on Θacc in Figure 1a) up to a maximum value of PBmax = 0.3 eV (as derived from our data of α-NPD on bare ITO): PBrem = (1 − Θacc)PBmax. The work function change upon α-NPD-deposition (ΔΦα−NPD) is the sum of the work function change due to pinning (ΔΦpin) and the remaining pushback (PBrem):

and 6.3 eV (F4TCNQ) is observed. For higher acceptor coverages Φsub remains constant. The saturation is attributed to the completion of a full monolayer, and acceptor multilayer formation does not further influence Φsub.14 Consequently, with each of the acceptor molecules we are able to access the range of interest around Φcrit = 4.8 eV, which is the expected transition value to drive α-NPD from vacuum level alignment to EF-pinning. To study the level alignment and the transition from vacuum level alignment to Fermi level pinning in detail, 5 nm thick αNPD films (corresponding to ∼5 layers) are deposited on acceptor precovered (from zero to one full monolayer) ITO substrates. Figure 1b displays the work function of such thick αNPD films (Φα−NPD) as well as its HOMO-onset position as a function of the effective work function Φsub of the acceptor precovered substrate. When depositing α-NPD on bare ITO, the α-NPD HOMO-onset is at 1.4 eV binding energy (BE) with a Φα−NPD of 3.9 eV. The sum of these values yields an IE of 5.3 eV for α-NPD. An increase of Φsub rigidly shifts the αNPD HOMO to lower BE until pinning sets in. For higher Φsub (corresponding to higher acceptor precoverages) the α-NPD HOMO-onset is pinned 0.5 eV below EF, with a final work function Φα−NPD of 4.8 eV. Both values correspond well to what can be found in the literature17 for depositing α-NPD on a multitude of different substrates (metals, metal oxides, conductive polymers) covering a wide interval of Φsub. However, upon further inspection of Figure 1 b, one finds that the Φcrit, at which pinning sets in in our experiments, is 5.2 eV for both acceptor molecules, even though it is expected to be equal to the α-NPD multilayer work function in the pinned regime, i.e., at 4.8 eV as noted in the Introduction. To explain this discrepancy of 0.4 eV, we involve a model that takes into account the nonuniformity of the acceptor precovered substrate (comprising coexisting covered and uncovered areas next to each other) as well as the pushback effect of α-NPD on bare ITO. In essence, this model considers the different local work function on bare ITO and acceptor-covered ITO surface patches, which together yield the surface averaged effective Φsub. As a prerequisite, we first discuss the two limiting cases of uniform surfaces, i.e., (i) zero and (ii) full acceptor coverage. (i) When depositing α-NPD on bare ITO with ΦITO of 4.2 eV, the work function is reduced by 0.3 eV (triangle, Figure 1b), which is attributed to the pushback-effect.18 (ii) For the acceptor monolayers (HATCN/F4TCNQ: Φsub = 5.8 eV/6.3 eV), the α-NPD HOMO-onset pins at 0.5 eV BE, and the work function is reduced due to interface dipoles by 1.0 eV/1.5 eV. For the intermediate effective substrate work function cases, i.e., α-NPD on submonolayers of acceptor molecules, some of the deposited α-NPD molecules adsorb on bare ITO as well as on top the acceptor molecules. The fraction of α-NPD molecules sticking directly on ITO cause pushback. The second process involved is electron transfer from those α-NPD molecules that reside on acceptor molecules, as the local work function there is already beyond that needed for pinning. For the relative contribution of both processes to the overall work function change, the growth mode governs the quantitative behavior. For instance, in the case of large coexisting domains of bare ITO and acceptor-covered areas, α-NPD molecules in the second layer would align either to the locally low ITOrelated work function or the local high work function related to acceptor-covered patches. As a result, one would observe a superposition of two differently aligned α-NPD species. Even if

ΔΦα‐NPD = ΔΦpin + PBrem

(1)

ΔΦpin at submonolayer acceptor coverages is the only not experimentally accessible parameter, but depends on the number of electrons Q transferred (due to pinning) from the α-NPD layer to the substrate, related via the Helmholtz equation: ΔΦpin =

Qεorg (2)

ed

where we use d = 0.3 nm as the thickness of one organic layer, e as the elementary charge, and εorg = 3.3522 as relative permittivity. At the premises of electronic equilibrium, the number of charges itself is given by Fermi-Dirac statistics: en Q= E 1 + exp k T

( ) B

(3)

where n is the number density of molecules per layer, T is the temperature, kB is the Boltzmann constant, and E is the final αNPD HOMO-onset position. As a last step, the final α-NPD HOMO-onset position is given by its initial position before adsorption (which is the effective work function of the substrate minus the α-NPD IE) minus the total work function change Δ Φα‑NPD: E = Φsub − IE − ΔΦα‐NPD

(4)

Also, for Φsub, the acceptor precoverage has to be taken into account with Φsub = ΦITO + θacc ΔΦacc, where ΦITO is the work function of bare ITO, and θaccΔΦacc is the coverage-dependent C



Article

Figure 2. (a) Evolution of the ΔΦα−NPD of a 5 nm thick α-NPD film as a function of Φsub. A constant shift of 0.3 eV persists until pinning is established. (b) Modeled (see text) evolution of ΔΦα‑NPD of a thick α-NPD film as a function of Φsub.

covered surface patches. The overall behavior of the work function is successfully described with an electrostatic model that accounts for both mechanisms that change the work function upon α-NPD deposition on a nonuniform substrate. The superposition of local work function values of dissimilar surface areas can fully explain the “delayed” pinning, which, with values of 0.4 eV in the present case, can be substantial. To ensure proper charge injection with molecularly modified electrodes, the lateral distribution of local work function values thus has to be considered to choose appropriate area-averaged work function values.

fraction of work function increase induced by a full acceptor monolayer (ΔΦacc). Inserting eq 4 into eq 3 and combining with eq 2, one can solve for ΔΦpin (and ΔΦα‑NPD likewise, according to eq 1) selfconsistently. en

(

1 + exp

Φsub − IE − (ΔΦpin + PBrem) kBT

)

=

ΔΦpinεorg ed (5)

In comparison to the corresponding experimental data in Figure 2a, the results of this model are plotted in Figure 2b for multilayer α-NPD as a function of θacc (which is also proportional to Φsub). We depict the contributions of PBrem and ΔΦpin separately. Increasing θacc (∼ Φsub), and therefore reducing the bare ITO area, leads to a linear decrease of ΔΦα‑NPD, which originates from the remaining pushback effect. It dominates as long as Φcrit is not reached. When the effective Φsub exceeds Φcrit, the pinning-induced interface dipole ΔΦpin starts to overcompensate the pushback effect and dominates for higher Φsub. The offset between the fundamental Φcrit and the effective Φsub at which pinning is observed experimentally (0.4 eV in the present case) can be reconciled by comparing Δ Φα‑NPD (including the pushback contribution) with ΔΦpin (no pushback contributions). This clearly shows that the origin of the delayed EF-pinning for acceptor submonolayer coverages is caused by the remaining push back effect of α-NPD. The competing effects that contribute to effective work function changes for nonuniform substrates must therefore carefully be included in the discussion and quantification of Fermi level pinning for organic semiconductor interfaces. For future work, it will be important to spatially resolve the electrode local work function variation, e.g., by scanning probe methods.

■

ASSOCIATED CONTENT

S Supporting Information *

Valence electronic structure of all employed materials within the heterostructure (ITO, HATCN, F4TCNQ, α-NPD), including a closer look at the additional photoelectron density found within the energy gap of the acceptor-molecules upon employing low Θacc indicating a charge-transfer-type interaction is provided. Together with the spectral shape variations of the α-NPD HOMO emission depending on Φsub (∼Θacc). This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Mailing address: Brook-Taylor-Straße 6, 12489 Berlin, Germany. Telephone number: +49 30 2093 7819. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

■

ACKNOWLEDGMENTS This work was supported by the DFG (SPP1355, SFB951) and the Helmholtz-Energie-Allianz “Hybrid-Photovoltaik”.

CONCLUSION The present photoemission study shows that Φsub can continuously be tuned between 4.2 and 6.3 eV as a function of strong molecular acceptor (HATCN and F4TCNQ) coverage on ITO. Accordingly, the HOMO-onset BE of αNPD deposited on top of such electrodes can be adjusted, and pinning of the HOMO-onset is found to be 0.5 eV below EF. This value was reached for a critical substrate work function Φcrit at 5.2 eV, which is in contrast to the commonly expected Φcrit value of 4.8 eV. This “delayed” pinning is rationalized by taking into account the pushback-effect induced by α-NPD deposited on bare ITO surface patches (before complete acceptor monolayer formation), which contributes to work function changes in addition to that occurring on acceptor-

REFERENCES

(1) Ishii, H.; Sugiyama, K.; Ito, E.; Seki, K. Energy Level Alignment and Interfacial Electronic Structures at Organic Metal and Organic Organic Interfaces. Adv. Mater. 1999, 11, 605−625. (2) Koch, N. Organic Electronic Devices and Their Functional Interfaces. Chem. Phys. Chem. 2007, 8, 1438−1455. (3) Braun, S.; Salaneck, W. R.; Fahlman, M. Energy-Level Alignment at Organic/Metal and Organic/Organic Interfaces. Adv. Mater. 2009, 21, 1450−1472. (4) Hill, I. G.; Rajagopal, A.; Kahn, A.; Hu, Y. Molecular Level Alignment at Organic Semiconductor−Metal Interfaces. Appl. Phys. Lett. 1998, 73, 662−664.

D



Article

(5) Witte, G.; Lukas, S.; Bagus, P. S.; Wöll, C. Vacuum Level Alignment at Organic/Metal Junctions: “Cushion” Effect and the Interface Dipole. Appl. Phys. Lett. 2005, 87, 263502. (6) Bagus, P. S.; Staemmler, V.; Wöll, C. Exchangelike Effects for Closed-Shell Adsorbates: Interface Dipole and Work Function. Phys. Rev. Lett. 2002, 89, 096104. (7) Bagus, P. S.; Hermann, K.; Wöll, C. The interaction of C6H6 and C6H12 with Noble Metal Surfaces: Electronic Level Alignment and the Origin of the Interface Dipole. J. Chem. Phys. 2005, 123, 184109. (8) Koch, N.; Kahn, A.; Ghijsen, J.; Pireaux, J. J.; Schwartz, J.; Johnson, R. L.; Elschner, A. Conjugated Organic Molecules on Metal versus Polymer Electrodes: Demonstration of a Key Energy Level Alignment Mechanism. Appl. Phys. Lett. 2003, 82, 70−72. (9) Hwang, J.; Wan, A.; Kahn, A. Energetics of Metal−Organic Interfaces: New Experiments and Assessment of the Field. Mater. Sci. Eng., R 2009, 64, 1−31. (10) Wan, A.; Hwang, J.; Amy, F.; Kahn, A. Impact of Electrode Contamination on the α-NPD/Au Hole Injection Barrier. Org. Electron. 2005, 6, 47−54. (11) Bröker, B.; Blum, R. P.; Beverina, L.; Hofmann, O. T.; Sassi, M.; Ruffo, R.; Pagani, G. A.; Heimel, G.; Vollmer, A.; Frisch, J.; Rabe, J. P.; Zojer, E.; Koch, N. A High Molecular Weight Donor for Electron Injection Interlayers on Metal Electrodes. Chem. Phys. Chem. 2009, 10, 2947−2954. (12) Osikowicz, W.; de Jong, M. P.; Salaneck, W. R. Formation of the Interfacial Dipole at Organic−Organic Interfaces: C60/Polymer Interfaces. Adv. Mater. 2007, 19, 4213−4217. (13) Yogev, S.; Matsubara, R.; Nakamura, M.; Zschieschang, U.; Klauk, H.; Rosenwaks, Y. Fermi Level Pinning by Gap States in Organic Semiconductors. Phys. Rev. Lett. 2013, 110, 036803. (14) Koch, N.; Duhm, S.; Rabe, J. P.; Vollmer, A.; Johnson, R. L. Optimized Hole Injection with Strong Electron Acceptors at Organic− Metal Interfaces. Phys. Rev. Lett. 2005, 95, 237601. (15) Bröker, B.; Blum, R. P.; Frisch, J.; Vollmer, A.; Hofmann, O. T.; Rieger, R.; Müllen, K.; Rabe, J. P.; Zojer, E.; Koch, N. Gold Work Function Reduction by 2.2 eV with an Air-Stable Molecular Donor Layer. Appl. Phys. Lett. 2008, 93, 243303. (16) Rana, O.; Srivastava, R.; Chauhan, G.; Zulfequar, M.; Husain, M.; Srivastava, P. C.; Kamalasanan, M. N. Modification of Metal− Organic Interface using F4-TCNQ for Enhanced Hole Injection Properties in Optoelectronic Devices. Physica Status Solidi A 2012, 209, 2539−2545. (17) Braun, S.; Osikowicz, W.; Wang, Y.; Salaneck, W. R. Energy Level Alignment Regimes at Hybrid Organic−Organic and Inorganic− Organic Interfaces. Org. Electron. 2007, 8, 14−20. (18) Greiner, M. T.; Helander, M. G.; Tang, W.-M.; Wang, Z.-B.; Qiu, J.; Lu, Z.-H. Universal Energy-Level Alignment of Molecules on Metal Oxides. Nat. Mater. 2012, 11, 76−81. (19) Koller, G.; Winter, B.; Oehzelt, M.; Ivanco, J.; Netzer, F. P.; Ramsey, M. G. The Electronic Band Alignment on Nanoscopically Patterned Substrates. Org. Electron. 2007, 8, 63−68. (20) Salzmann, I.; Duhm, S.; Heimel, G.; Oehzelt, M.; Kniprath, R.; Johnson, R. L.; Rabe, J. P.; Koch, N. Tuning the Ionization Energy of Organic Semiconductor Films: The Role of Intramolecular Polar Ponds. J. Am. Chem. Soc. 2008, 130, 12870−12871. (21) Ley, L.; Smets, Y.; Pakes, C. I.; Ristein, J. Calculating the Universal Energy-Level Alignment of Organic Molecules on Metal Oxides. Adv. Funct. Mater. 2013, 23, 794. (22) Noguchi, Y.; Miyazaki, Y.; Tanaka, Y.; Sato, N.; Nakayama, Y.; Schmidt, T. D.; Brütting, W.; Ishii, H. Charge Accumulation at Organic Semiconductor Interfaces due to a Permanent Dipole Moment and Its Orientational Order in Bilayer Devices. J. Appl. Phys. 2012, 111, 114508.

E


The Impact of Local Work Function Variations on Fermi Level Pinning

Recommend Documents