THE JOURNAL OF CHEMICAL PHYSICS 142, 184702 (2015)

An electron energy-loss study of picene and chrysene based charge transfer salts Eric Müller, Benjamin Mahns, Bernd Büchner, and Martin Knupfer IFW Dresden, P.O. Box 270116, D-01171 Dresden, Germany

(Received 12 February 2015; accepted 27 April 2015; published online 8 May 2015) The electronic excitation spectra of charge transfer compounds built from the hydrocarbons picene and chrysene, and the strong electron acceptors F4TCNQ (2,3,5,6-tetrafluoro-7,7,8,8tetracyanoquinodimethane) and TCNQ (7,7,8,8-tetracyanoquinodimethan) have been investigated using electron energy-loss spectroscopy. The corresponding charge transfer compounds have been prepared by co-evaporation of the pristine constituents. We demonstrate that all investigated combinations support charge transfer, which results in new electronic excitation features at low energy. This might represent a way to synthesize low band gap organic semiconductors. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4919881]

I. INTRODUCTION

As a consequence of the size of their band gap, π-conjugated organic materials can be typical insulators or semiconductors, respectively. The latter, in general, are in the focus of both fundamental and applied researches. Their possible application in organic electronic devices ranges from fieldeffect transistors1 over light-emitting diodes2,3 to organic solar cells.4 Moreover, due to their relatively open crystal structure, the electronic properties can be easily modified by the addition of electron acceptors and donors, which enables to tune their physical behaviour by the addition or removal of charges to or from the molecules. Prominent examples comprise the enhanced efficiency of organic devices because of doping,5,6 the appearance of superconductivity in alkali metal doped molecular solids (e.g., fullerides7–9 and picene10), as well as the formation of charge transfer salts11 or of a two dimensional metallic interface between two organic insulators (7,7,8,8tetracyanoquinodimethane (TCNQ) and TTF12).13 Moreover, various charge transfer complexes consisting of typical organic semiconductors and strong electron acceptors have been discussed recently in the context of their promising characteristics for high-mobility ambipolar charge transport.14–17 This would open up a new route to materials with high potential for organic electronic. Also, it has been demonstrated that charge transfer salt single crystals out of picene and the strong electron acceptor 2,3,5,6-tetrafluoro-7,7,8,8tetracyanoquinodimethane (F4TCNQ) can be grown using physical vapor transport.18 Furthermore, transport studies have indicated an activated transport with energy gaps much smaller than those of the starting materials. Thus, this might represent a route to small band gap organic semiconductors, an further extension of this semiconducting material class. We have extended the previous results combining the hydrocarbons picene and chrysene with the strong electron acceptors F4TCNQ and TCNQ. This allows to shine light on the impact of the size of the molecules and their energy of their electronic levels on the resulting charge transfer. Moreover, in consideration of recent reports on superconductivity in 0021-9606/2015/142(18)/184702/4/$30.00

electron doped picene,10 our studies also complement previous work on potassium doped picene and chrysene19,20 in the direction of molecular crystals with positively charged picene and chrysene. The latter consists of four and five benzene rings, respectively, which are connected in a zigzag manner. The molecular structures of these materials are depicted in Figure 1. Our studies provide further insight into the electronic properties of recently grown F4TCNQ-based single crystals and furthermore yield additional knowledge on the reaction of TCNQ with hydrocarbons. We demonstrate that for all combinations, new electronic excitations show up in the band gap of the pure starting materials. Moreover, our analysis of the infra-red (IR) measurements shows a partial charge transfer in all cases.

II. EXPERIMENT

Thin films composed of an electron acceptor (TCNQ and F4TCNQ) and an electron donor material (picene and chrysene) were prepared by thermal co-evaporation under high vacuum onto single crystalline KBr substrates kept at room temperature. The evaporators for the acceptor and the donor material are oriented in a way that the two materials are mixed in the gas phase, before they reach the substrate. The deposition rate was about 0.5 nm/min for each material to achieve a stoichiometry close to 1:1. In particular, for the F4TCNQ/picene mixed film, single crystals of F4TCNQ/picene18 were evaporated to achieve an ideal stoichiometry. For all combinations, the film thickness was about 100 nm. These films were then floated off in distilled water, mounted onto standard electron microscopy grids, and transferred into the spectrometer, where we investigated the electronic excitation spectrum using electron energy-loss spectroscopy (EELS) in transmission. This technique has provided valuable insight into the electronic excitations of molecular solids in the past (e.g., Refs. 21–23). All electron diffraction studies and excitation measurements were carried out using the 172 keV spectrometer described in detail elsewhere.24,25 The energy and momentum resolution

142, 184702-1

© 2015 AIP Publishing LLC

184702-2

Müller et al.

J. Chem. Phys. 142, 184702 (2015)

FIG. 1. Schematic representation of the molecular structure of the investigated hydrocarbons.

were chosen to be 82 meV and 0.04 Å−1. We have measured the loss function Im[−1/ϵ(q,ω)], which is proportional to the dynamic structure factor S(q,ω), for a momentum transfer q parallel to the film surface [ϵ(q,ω) is the dielectric function]. Infrared measurements were taken using a Bruker interferometer spectrometer 88 spectrometer with a spectral resolution of 2 cm−1 in a range of 400–4000 cm−1. The signal was counted by a deuterated Triglycine sulfate (DTGS)-detector; as light source, a globar was used. The spectra have directly been taken on the thin films prepared on KBr as described above. Measuring the infrared absorption bands of the C≡≡N triple bond of F4TCNQ in the charge transfer compounds with picene and chrysene, respectively, reveals a shift to lower wave numbers with respect to the neutral F4TCNQ as shown in Fig. 2. This shift signals a charge transfer in the compounds as has been discussed previously for F4TCNQ/picene single

FIG. 3. Elastic electron diffraction spectra of the F4TCNQ/picene thin films with the corresponding peak positions resulting from X-ray diffraction of respective single crystals.18 We note that our momentum resolution is much less than that of an X-ray diffraction experiment, which renders it impossible to identify very close lying diffraction peaks.

crystals in detail18 and the size of this charge transfer has been determined to be 0.14–0.19 electrons. Concerning the TCNQ compounds, the evaluation of the infrared absorption bands results in a charge transfer in the same range. Moreover, comparing the positions of the C≡≡N absorption band of F4TCNQ/picene thin film (2223 cm−1) with the position in F4TCNQ/picene single crystals (2222 cm−1),18 we see a very good correspondence. Equivalently, there is also a very good agreement for the respective line positions for the F4TCNQ/chrysene thin film (2223 cm−1) and the F4TCNQ/chrysene single crystals (2222 cm−1).27,28 We thus conclude that our thin films are representative of the corresponding charge transfer crystals, and their electronic excitation spectra represent those of the charge transfer compounds. This conclusion is fully corroborated by the electron diffraction profiles which have been measured using the EELS for all compounds, e.g., see Fig. 3. They indicate that our thin films are polycrystalline, and we were able to rationalize all observed diffraction peaks for the F4TCNQ/picene and F4TCNQ/chrysene films on the basis of the known crystal structure.18,28 This is illustrated in Figure 3 for the case of F4TCNQ/picene.

III. RESULTS AND DISCUSSION

FIG. 2. Comparison of the IR absorption band of the C≡≡N triple bond in pure F4TCNQ and in charge transfer complexes with picene and chrysene, respectively, as prepared and investigated in this contribution. The shift to lower wave numbers signals a charge transfer of about 0.14–0.19 electrons per F4TCNQ.18,26

In Fig. 4, we show the loss function of F4TCNQ/picene in comparison to those of the two starting materials. The measurements have been carried out with a momentum transfer of 0.1 Å−1, i.e., they represent the optical limit and the results are thus comparable to optical data. The inset of Fig. 4 depicts a comparison of the loss function of F4TCNQ/picene thin films as measured using EELS (solid red line) to the optical absorption spectrum of single crystalline F4TCNQ/picene (dashed black line). Apart from small energy shift and slightly different

184702-3

Müller et al.

J. Chem. Phys. 142, 184702 (2015)

FIG. 4. Comparison of the loss function of a F4TCNQ/picene film with the spectra of the pure materials. The measurements have been carried out with a momentum transfer of 0.1 Å−1. The inset shows a comparison of the EELS loss function of F4TCNQ/picene films (solid red line) to the optical absorption spectrum of single crystals (dashed black line). Note that the two techniques measure different response functions, which can cause slightly different energy positions in the respective spectra.

broadening, the two curves agree very well. As a consequence, our EELS data represent well the electronic excitation spectrum of the corresponding single crystalline material as were already indicated by the IR and diffraction results discussed above. The spectrum of pristine picene is characterized by an excitation onset at 3.1 eV which represents the lowest singlet exciton and which is followed by further excitons and interband transitions between occupied and unoccupied energy levels.23,29 The loss function of F4TCNQ also shows a relatively large gap with an excitation onset at about 2.6 eV, which is in good agreement to previous optical studies.30 In the energy window below the excitation onset of both starting materials, the F4TCNQ/picene film reveals an excitation centered at 1.84 eV with an excitation onset of 1.35 eV. This feature is followed by a shoulder at about 2.3 eV and an additional prominent excitation at 2.9 eV. All these new excitation features cannot be the result of a superposition of the loss function of the starting materials, and thus, they represent clear evidence for the formation of a charge transfer compound between the electron acceptor F4TCNQ and the electron donor picene. Following a previous analysis of F4TCNQ/picene dimers using density functional based calculations,18 a hybrid orbital is formed out of the HOMO-1 (HOMO, highest occupied molecular orbital) orbital of picene and the lowest unoccupied molecular orbital (LUMO) of F4TCNQ. Interestingly, this hybrid orbital is the HOMO-1 state of the dimer. This interaction also results in a charge transfer of about 0.14–0.19 electrons as revealed by the calculations and experimental analyses.18 The new low energy electronic excitations are thus a consequence of the formation of the charge transfer com-

FIG. 5. Comparison of the loss function of F4TCNQ/picene and TCNQ/picene (panel (a)), and of F4TCNQ/chrysene and TCNQ/chrysene (panel (b)). The measurements have been carried out with a momentum transfer of 0.1 Å−1.

pound. In other words, charge transfer complexes on the basis of hydrocarbons and the electron acceptor F4TCNQ can be regarded as one possibility to achieve organic crystals with relatively low band gaps. In the following, we present data on three other charge transfer materials based upon the related hydrocarbons picene and chrysene, and F4TCNQ and its unfluorinated relative TCNQ. In Fig. 5 (panel (a)), we compare the loss function of TCNQ/picene and that of F4TCNQ/picene. Clearly, both materials support electronic excitations at energies below 3 eV which are not observed for the respective starting materials. The first two excitation features for TCNQ/picene at energy positions of 2 eV and 3 eV are only slightly shifted to higher energies compared to F4TCNQ/picene, which is also the case for the spectral onset at about 1.5 eV. This difference most

184702-4

Müller et al.

likely arises from the variation of the electron affinity going from TCNQ to F4TCNQ. For energies above 3.5 eV, the excitation spectra are quite similar. Moreover, equivalent spectra and changes upon charge transfer compound formation are observed for TCNQ/chrysene and F4TCNQ/chrysene, also shown in Fig. 5 (panel (b)). Again, new electronic excitations show up below the excitation onset of the starting materials (see, e.g., Fig. 4 and Ref. 19). In contrast to the picene compounds, low energy shoulders appear at about 1.65 eV and 1.9 eV, next to the first excitation peaks at 2.15 eV and 2.45 eV for F4TCNQ/chrysene and TCNQ/chrysene, respectively. These features are shifted to higher energies for the F4TCNQ based material, analogous to the picene compounds. The spectral onset is found at 1.4 eV for F4TCNQ/chrysene and at 1.5 eV for TCNQ/chrysene. Above 2.6 eV, we again observe quite similar excitation spectra in the chrysene compounds. Consequently, our data indicate a clear trend in regard of the electronic excitation spectra of charge transfer compounds built from phenacene hydrocarbons and TCNQ or F4TCNQ. The charge transfer between the molecules is responsible for new electronic excitations below the excitation onset of the two starting materials. In either case, the reaction with the F4TCNQ molecule, characterized by the stronger electron affinity, results in the energetically lowest excitation onset and excitation features. On the other hand, going from picene to chrysene, the onset is hardly changing while the spectral shape is rather different with low energy shoulders below the first maximum in the case of the chrysene compounds. Finally, it is interesting to realize that previous investigations of the electrical transport behaviour of F4TCNQ/picene single crystals18 have revealed an activated behaviour with an activation energy of about 0.6 eV. This value is much lower in energy than the spectral onset in our loss function (or in optical absorption data) found at 1.35 eV. This difference might be related to the nature and symmetry of the HOMO and LUMO of the F4TCNQ/picene compound, which results in an optically forbidden gap excitation in this material. Indeed, the calculations of a F4TCNQ-picene dimer18 support this scenario.

IV. SUMMARY

To summarize, we have investigated the electronic excitations of co-evaporated films of the electron acceptors TCNQ and F4TCNQ combined with the electron donor materials chrysene and picene using electron energy-loss spectroscopy in transmission. An analysis of additional IR and diffraction data demonstrates that our films are representative of the corresponding phase pure compound. We have shown that in either case new electronic excitations appear in the gap of the two starting materials, and the spectral onset is always observed slightly below the visible region. This might point towards a route which allows to grow organic semiconducting materials with relatively low band gaps. In addition, these materials might also support ambipolar charge transport,14–17 and our

J. Chem. Phys. 142, 184702 (2015)

studies can thus be regarded as an important characterization of prospective organic electronic materials. ACKNOWLEDGMENTS

We thank M. Naumann, R. Hübel, and S. Leger for technical assistance. This work has been supported by the Deutsche Forschungsgemeinschaft (Grant No. KN393/14). 1H.

Okamoto, N. Kawasaki, Y. Kaji, Y. Kubozono, A. Fujiwara, and M. Yamaji, J. Am. Chem. Soc. 130, 10470 (2008). 2S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lussem, and K. Leo, Nature 459, 234 (2009). 3M. Gross, D. C. Muller, H.-G. Nothofer, U. Scherf, D. Neher, C. Brauchle, and K. Meerholz, Nature 405, 661 (2000). 4M. Niggemann, B. Zimmermann, J. Haschke, M. Glatthaar, and A. Gombert, Thin Solid Films 516, 7181 (2008). 5K. Walzer, B. Maennig, M. Pfeiffer, and K. Leo, Chem. Rev. 107, 1233 (2007). 6J. Blochwitz, M. Pfeiffer, T. Fritz, and K. Leo, Appl. Phys. Lett. 73, 729 (1998). 7O. Gunnarsson, Rev. Mod. Phys. 69, 575 (1997). 8O. Gunnarsson, J. Han, E. Koch, and V. Crespi, Superconductivity in Complex Systems, Structure and Bonding Vol. 114, edited by K. Müller and A. Bussmann-Holder (Springer, Berlin, Heidelberg, 2005), pp. 71–101. 9J. Weaver and D. Poirier, Solid State Physics, Solid State Physics Vol. 48, edited by H. Ehrenreich and F. Spaepen (Academic Press, 1994), pp. 1–108. 10R. Mitsuhashi, Y. Suzuki, Y. Yamanari, H. Mitamura, T. Kambe, N. Ikeda, H. Okamoto, A. Fujiwara, M. Yamaji, N. Kawasaki, Y. Maniwa, and Y. Kubozono, Nature 464, 76 (2010). 11N. Toyota, M. Lang, and J. Müller, Low-Dimensional Molecular Metals, Springer Series in Solid-State Sciences Vol. 154 (Springer, Berlin, 2010). 12Tetrathiafulvalene. 13H. Alves, A. S. Molinari, H. Xie, and A. F. Morpurgo, Nat. Mater. 7, 574 (2008). 14S. Horiuchi, T. Hasegawa, and Y. Tokura, J. Phys. Soc. Jpn. 75, 051016 (2006). 15K. P. Goetz, D. Vermeulen, M. E. Payne, C. Kloc, L. E. McNeil, and O. D. Jurchescu, J. Mater. Chem. C 2, 3065 (2014). 16L. Zhu, Y. Yi, Y. Li, E.-G. Kim, V. Coropceanu, and J.-L. Brédas, J. Am. Chem. Soc. 134, 2340 (2012). 17L. Zhu, Y. Yi, A. Fonari, N. S. Corbin, V. Coropceanu, and J.-L. Brédas, J. Phys. Chem. C 118, 14150 (2014). 18B. Mahns, O. Kataeva, D. Islamov, S. Hampel, F. Steckel, C. Hess, M. Knupfer, B. Büchner, C. Himcinschi, T. Hahn, R. Renger, and J. Kortus, Cryst. Growth Des. 14, 1338 (2014). 19F. Roth, B. Mahns, R. Schönfelder, S. Hampel, M. Nohr, B. Büchner, and M. Knupfer, J. Chem. Phys. 137, 114508 (2012). 20F. Roth, B. Mahns, B. Büchner, and M. Knupfer, Phys. Rev. B 83, 144501 (2011). 21M. Knupfer, T. Pichler, M. S. Golden, J. Fink, M. Murgia, R. H. Michel, R. Zamboni, and C. Taliani, Phys. Rev. Lett. 83, 1443 (1999). 22R. Schuster, M. Knupfer, and H. Berger, Phys. Rev. Lett. 98, 037402 (2007). 23F. Roth, P. Cudazzo, B. Mahns, M. Gatti, J. Bauer, S. Hampel, M. Nohr, H. Berger, M. Knupfer, and A. Rubio, New J. Phys. 15, 125024 (2013). 24J. Fink, Recent Developments in Energy-Loss Spectroscopy, Advances in Electronics and Electron Physics Vol. 75, edited by P. W. Hawkes (Academic Press, 1989), pp. 121–232. 25F. Roth, A. König, J. Fink, B. Büchner, and M. Knupfer, J. Electron Spectrosc. Relat. Phenom. 195, 85 (2014). 26J. S. Chappell, A. N. Bloch, W. A. Bryden, M. Maxfield, T. O. Poehler, and D. O. Cowan, J. Am. Chem. Soc. 103, 2442 (1981). 27M. Nohr, “Einkristallzucht und Charakterisierung cyclischer Ladungstransfersalze,” M.S. thesis (University of Applied Sciences/Leibniz Institute for Solid State and Materials Research Dresden, 2014). 28M. Nohr, “Growth and characterization of chrysene-F4TCNQ and naphthacene-F4TCNQ charge transfer salts” (unpublished). 29F. Roth, B. Mahns, B. Büchner, and M. Knupfer, Phys. Rev. B 83, 165436 (2011). 30D. A. Dixon, J. C. Calabrese, and J. S. Miller, J. Phys. Chem. 93, 2284 (1989).

Journal of Chemical Physics is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.