Orbital-Resolved Partial Charge Transfer from the Methoxy Groups of Substituted
Pyrenes in Complexes with Tetracyanoquinodimethane
- a NEXAFS Study
K. Medjanik, S. A. Nepijko, H. J. Elmers, G. Schönhense*
Institut für Physik, Johannes Gutenberg-Universität Mainz, Germany
P. Nagel, M. Merz, S. Schuppler
Karlsruhe Institute of Technology (KIT), Institut für Festkörperphysik, 76021 Karlsruhe,
D. Chercka, M. Baumgarten, and K. Müllen
Max-Planck-Institut für Polymerforschung, Mainz, Germany
It is demonstrated that the near-edge X-ray absorption fine structure (NEXAFS) provides a
powerful local probe of functional groups in novel charge transfer (CT) compounds.
Microcrystals of tetra- and hexamethoxypyrene as donors with the strong acceptor
tetracyanoquinodimethane (TMPx/HMPx - TCNQy) were grown from solution via vapour
diffusion in different stoichiometries x:y = 1:1, 1:2 and 2:1. Owing to the element specificity
of NEXAFS, the oxygen and nitrogen K-edge spectra are direct spectroscopic fingerprints of
the donating and accepting moieties. The orbital selectivity of the NEXAFS resonances
allows to precisely elucidate the participation of specific orbitals in the charge-transfer
process. In the present case charge is transferred from methoxy-orbitals 2e (π*) and 6a1(σ*)
to the cyano-orbitals b3g and au (π*) and - to a weaker extent - to b1g and b2u (σ*). The
occupation of 2e reflects the anionic character of the methoxy groups. Surprisingly, the charge
transfer increases with increasing HMP content of the complex. As additional indirect
signature, all spectral features of the donor and acceptor are shifted to higher and lower
photon energies, respectively. Providing quantitative access to the relative occupation of
specific orbitals, the approach constitutes the most direct probe of the charge-transfer
mechanism in organic salts found so far. Although demonstrated for the specific example of
pyrene-derived donors with the classical acceptor TCNQ, the method is very versatile and can
serve as routine probe for novel CT-complexes on the basis of functionalized polycyclic
1.1 Functionalized Polycyclic Aromatic Hydrocarbons With Strong Acceptor or Donor
In the field of molecular electronics conjugated organic molecules have received intense
attention as n-type or p-type semiconductors. In particular, charge transfer (CT) compounds
of molecules with tailored donor and acceptor character provide a vast multitude of design
possibilities. Understanding the electronic structure of this class of materials as well as their
metal-organic interfaces is crucial for designing specific electrical properties. In particular,
large planar polycyclic aromatic hydrocarbon molecules with different functional groups at
the periphery have recently been studied intensively. It was shown that their electronic
properties can be tailored for strong electron acceptor or donor character . Novel chemical
synthesis routes as described for the coronene case by Rieger et al.  paved the way towards
the design of a new class of donor and acceptor molecules both based on the same parent
molecule. Indeed, the UHV co-deposited donor hexamethoxycoronene and acceptor
coronene-hexaone form a weak CT complex as recently shown using photoemission and
infrared spectroscopy .
Likewise, the pyrene molecule can be functionalized at its periphery by adding either
methoxy- or keto-groups thus yielding moderate donors or strong acceptors, respectively .
The donor moieties are of particular interest because their sizes are similar to that of the
classical strong acceptor 7,7,8,8-tetracyano-p-quinodimethane, (TCNQ, C12N4H4), see Fig.1a.
In 4,5,9,10-tetramethoxypyrene (TMP, C20H18O4) the four functional groups lower the
ionization potential (IP) to 5.47 eV , indicating a strongly increased donor character in
comparison with the parent molecule
2,4,5,7,9,10-hexamethoxypyrene (HMP, C22H22O6) IP is further lowered to 5.17 eV, as
measured by cyclovoltammetry. It is thus interesting to investigate the possible formation of
new CT complexes based on these donor moieties, in particular in compounds with TCNQ.
For the present study, new crystallographic phases with different stoichiometries of TCNQ
and HMP as well as TMP have been grown from solution.
Thin films of the TMP-TCNQ complex on atomically clean gold surfaces have been produced
by UHV co-deposition and investigated by ultraviolet photoelectron spectroscopy (UPS) and
scanning tunnelling spectroscopy (STS) . New reflexes in X-ray diffraction from these
films gave evidence of a crystallographic phase different from pure TCNQ and TMP. The
reflex positions in θ-2θ scans of the thin film samples are compatible with the results of the
3D-analysis of the solution-grown crystallites. Infrared spectra revealed a red-shift of the CN
stretching vibration frequency by 7 cm-1, indicating a charge transfer of about 0.3e. Shifts in
the level positions of the frontier orbitals as visible in UPS and STS could be qualitatively
interpreted on the basis of density functional theory calculations. The mixed-stack phases of
the pyrene-derived donors TMP and HMP with TCNQ thus constitute a new class of charge
Despite of the first results on the crystallographic structure of the solution-grown crystallites
 and the mainly spectroscopic results on the UHV-deposited thin films , several
important questions still remain unsolved. In particular, there is lack of information on the
electronic structure of the solution-grown crystals because the unavoidable surface
contaminations and the small size of the crystallites were prohibitive for UPS and STS
analyses. The probing depth is only 2-3 molecular layers in UPS and only one molecular layer
in STS. In the present work, we have employed X-ray absorption spectroscopy. It provides a
pyrene (IP = 7.41 eV ). In
larger probing depth in the order of 5 nm. Thus, surface contaminations are expected to be
much less severe than in UPS or STS.
1.2 Charge Transfer Salts Studied via Near-Edge X-ray Absorption Fine Structure
In NEXAFS measurements an electron is excited from a core level to an empty or partially
unoccupied valence electronic state. As this technique gives direct access to the unoccupied
density of states, it should be sensitive to the formation of new hole states due to charge
transfer in a donor-acceptor complex . By selection of a specific atom via its X-ray
absorption edge, the electronic structure in the vicinity of this atom (e.g. in a functional group
attached to a large molecule) is probed. For the present experiment the nitrogen 1s and
oxygen 1s core levels should be ideal, because these atomic species are located exclusively on
the acceptor site (N in the cyano-groups of TCNQ) or the donor site (O in the methoxy-groups
of TMP and HMP). This highly specific excitation should provide information on the local
electronic structure and thus on local chemical functionalities. We cannot expect carbon K-
edge NEXAFS to be particularly useful due to the many non-equivalent C-atoms in both the
donor and acceptor ring systems. Nitrogen and oxygen K-edge NEXAFS require
monochromatic, tuneable synchrotron radiation in the soft X-ray range. In the present study
we exploited soft X-rays from the WERA beamline at ANKA, Karlsruhe. In the total electron
yield mode employed, the drain current from the sample is detected. The electron emission
yield originates from a subsequent Auger process that neutralizes the core hole and leads to
the emission of Auger electrons and slow secondary electrons. The information depth is about
5 nm in total yield mode.
The NEXAFS method is described in detail in . Fraxedas et al.  performed a NEXAFS
study of the classical charge transfer salt tetrathiafulvalene (TTF, C6S4H4) –TCNQ. This
compound forms parallel segregated stacks of donors (TTF) and acceptors (TCNQ). Results
have been compared with a first-principles calculation of the unoccupied and partially
occupied electronic states of the pure materials and the charge transfer compound. Later, Sing
et al.  studied the same system with particular focus on the renormalized band widths
observed in UPS [11,12] for the same compound. By variation of the angle of photon
incidence, the symmetry of the observed orbitals for TCNQ was probed and information on
molecular orientation was gained . None of these papers addressed the issue of a possible
change of the unoccupied density of states upon formation of the CT complex.
In the present paper we present NEXAFS results for solution-grown 3D crystallites of the
complexes HMP-TCNQ and TMP-TCNQ in different stoichiometric mixtures. For
comparison, spectra of the pure donors and acceptors have been taken under the same
conditions. The study revealed two different signatures of a charge transfer in these new
compounds. Strong changes in the intensity of the oxygen pre-edge features for different
stoichiometries are a fingerprint of the occurrence of additional hole states at the donor sites.
This pre-edge feature is absent in the spectrum of pure donor material, giving evidence of full
anionicity of the methoxy group in HMP and TMP. Complementary, two prominent
resonances of TCNQ are quenched, indicating partial filling of these states. In addition,
characteristic energy shifts of the donor and acceptor NEXAFS resonances in opposite
directions are a further consequence of the charge transfer.
2. Crystal Growth of CT Compounds from Solution via Vapour Diffusion
In 1962 Melby et al.  have grown pyrene-TCNQ in solution and found a weak CT
complex; later Amano et al.  have grown diaminopyrene-TCNQ in acetonitrile CH3CN
and found an ionic CT salt. In first experiments we just looked at the colour change of
concentrated solution of the donor-acceptor moieties providing a broad CT absorption band in
the visible range with maximum at about 600 nm.
Crystals were grown by vapour diffusion of hexane into a dichloromethane solution (5 ml,
6,2*10-3 mol/l) of the components. Solutions with donor-acceptor mixtures of 1:2, 1:1 and 2:1
stoichiometry were prepared. The components were combined in a glas vial (V = 7 cc, 1,5 cm
diameter) and dissolved under sonication. Vapour diffusion assisted crystallizations
were performed in a gas-tight chamber (V = 120 cc), filled with 15 ml hexane. The
vial containing the solution was placed inside the chamber which was sealed for 4 days.
The crystallites have sizes in the range from several 10µm to several 100µm. X-ray
diffraction analysis of TMP-TCNQ microcrystal fractions revealed a mixed-stack geometry as
shown in Fig. 1b.  The complex appears black, whereas crystallites of the pure donors
HMP or TMP are colourless transparent and TCNQ is transparent with light green colour.
Optical microscopy revealed that there is a mixture of dark and transparent crystals in the vial,
examples, see Fig. 1c and d. The admixture of bright crystals is maximum in the HMP2-
TCNQ1-stoichiometry. For the 1:1 stoichiometry the admixture is smaller and for the 1:2
stoichiometry there are no transparent crystals. The different phases could easily be
distinguished by their colour and the crystal fractions could thus be separated using a
micromanipulator under the optical microscope. The NEXAFS spectra revealed that the
transparent crystals show no nitrogen K-edge signal. This proves that they contain no TCNQ
and consist of pure donor material. Obviously, HMP or TMP microcrystals form during the
vapour diffusion process in coexistence with the complex.
3. Nitrogen and Oxygen K-edge NEXAFS Spectra
NEXAFS spectra of the different complexes and of pure donor and acceptor molecules have
been taken at the WERA beamline of ANKA. The dark fractions of the solution-grown
microcrystals were deposited on carbon tape, being a suitable holder for such samples. The
results for the nitrogen and oxygen K-edge spectra are summarized in Fig. 2, 3, and 5. The
spectral features (A-E for nitrogen and F-K for oxygen) were quantitatively analyzed by a
multi-peak fit routine. The partial spectra resulting from the fit are shown as thin lines, circles
denote the measured spectra.
As nitrogen is present only in the cyano-groups of TCNQ, its edge fine structure is a
fingerprint of the acceptor. Likewise, oxygen is only contained in the methoxy-groups and
thus its spectrum represents a local probe in the functional group of the donor. The abscissa is
identical for all spectra shown. Clearly, the absolute total yield of the N K-edge spectra grows
with increasing TCNQ content of the compounds, cf. sequence of spectra a, b, c and d, in
Fig.2, the latter taken for pure TCNQ. Vice versa, the intensity of the oxygen features (Fig. 3)
decreases in the sequence of spectra e (taken for pure HMP), f, g and h. This proves that the
relative donor-acceptor concentrations in the solution show up in the stoichiometry of the 3D
crystallites as well.
Figure 1: Molecular structures (a) and structure of the TMP1-TCNQ1complex as obtained from
X-ray diffraction (b); the second plot shows a colour-coded top view of the molecular
arrangement in four adjacent layers. Optical microscopy of HMP2-TCNQ1(c) and TMP1-TCNQ1
(d) shows coexistence of dark and bright microcrystals (field of view 1mm horiz.).
3.1 Nitrogen K-edge NEXAFS
The spectrum for the nitrogen K-edge was deconvoluted into signals A-E using the multi-peak
fit routine, assuming a mixed Gaussian/Lorentzian function accounting for a finite energy
resolution of 0.2 eV and a life-time broadening of 0.4 eV. Fit results are plotted in Fig.2 as
thin curves below the spectra. The sum of all partial spectra is shown as full curve through the
data points (circles). In all cases the fit curve perfectly reproduces the data points, giving
evidence of a high reliability of the partial spectra. NEXAFS spectra of pure TCNQ have been
analyzed in Ref. . Here we briefly recall the peak assignment. The first signal (A) is a π*–
type resonance (resonant transition 1s→π*) and originates from the lower auand b1u orbitals.
For the free molecule these lie at 2.55 eV and 2.65 eV above the LUMO minimum. At photon
energies in resonance with such an allowed dipole transition, a large increase in excitation
cross section is observed. The orbitals auand b1u originate from the degenerate pair of lowest
empty π*-orbitals (e2u) of the benzene core, which only slightly delocalize towards the cyano-
group in TCNQ. We denote transition A as N1s→au,b1u[π π π π*(ring)]. Signals B and C originate
from p-type unoccupied orbitals located in the cyano-group and are therefore higher in
intensity due to the larger overlap with the N1s wavefunction in the dipole matrix element. B
is a σ*-type resonance originating from the b1g and b2u orbitals that belong to the four
symmetry adapted combinations of in-plane orbitals of the CN groups. The corresponding
transition B is denoted as N1s→b1g,b2u [σ σ σ σ*(C≡N)]. Signal C is of π*-type and derives from
the b3g and auorbitals, i.e. N1s→b3g,au[π π π π*(C≡N)]. Finally, signal D corresponds to the highest
π*-type orbital of benzene (b2g) that is delocalized over the whole TCNQ molecule,
N1s→b2g[π π π π*(ring)]. In summary, the unoccupied states involved in the weaker transitions A
and D are essentially delocalized on the benzene ring, whereas states involved in transitions B
and C largely consist of the σ*- and π*-orbitals of the cyano-group. The weak feature E and
higher lying weak signals (not shown) originate from delocalized σ-type orbitals containing
σ*(C-C), σ*(C-H) and σ*(C≡N) contributions.
Unlike the TCNQ-TTF case , where signals B and C merge to a single intense peak, they
can be well separated by the fit (Fig. 2 a – c). In spectrum (d), showing the result for pure
TCNQ powder, signal B appears as a low-energy shoulder. We have checked the order of
signals B and C by taking spectra for thin-film samples of TCNQ (not shown here): At
grazing incidence (75°) the π*-type signal (C) gains intensity, whereas at normal incidence
(0°) signal (B) with σ*-symmetry has its intensity maximum. This is in agreement with the
angular-dependent measurements of Sing et al. .
With increasing HMP content (sequence d, c, b, a) the signal intensities B and C drop
substantially. For HMP2-TCNQ1(spectrum a) signals B and C appear as a double peak with a
separation of about 190 meV. This intensity drop is an indication of charge being transferred
into these unoccupied orbitals of the acceptor molecule. Moreover, significant shifts of the
resonance positions towards lower hν occur as function of the HMP content (see vertical
lines). We will return to these points in the context of a complementary behaviour found for
the oxygen pre-edge signal of the donor molecule. For the TTF-TCNQ complex, only small
variations in N K-edge NEXAFS spectra in comparison with pure, neutral TCNQ have been
Figure 2: Nitrogen K-edge NEXAFS spectra of HMPx-TCNQy complexes with different
stoichiometries. The yield scale is the same for all spectra; spectrum d was scaled by a factor of
0.5. Circles denote experimental data; thin curves below the spectra mark partial spectra of the
transitions as obtained from a multi-peak fit routine; curves through the dots are the result of the
fit (sum of the partial spectra). Bottom panel: Same for TMP2-TCNQ1.
3.2 Oxygen K-edge NEXAFS
Oxygen is only contained in the methoxy-groups of TMP and HMP. So its edge fine structure
is a fingerprint of the donor moieties. Fig. 3 shows a series of spectra with f, g and h
corresponding to the same samples as the series a, b and c in Fig. 2. Spectrum e has been
taken for pure HMP. A number of signals (F-K) can be identified that partly show strong
intensity changes for different stoichiometries.
The oxygen K-edge NEXAFS of the methoxy species chemisorbed on transition metal
surfaces has been analyzed by Amemiya et al., we recall the peak assignments from . The
prominent pre-edge peak in the spectra (F) is separated by almost 3 eV from the next peak.
This peak is located at 532.2 eV, in good agreement with the position of the lowest-lying peak
measured for the methoxy species on Cu (531.7 eV ). This low-lying signal derives from
the highest occupied molecular orbital (HOMO) 2e of the methoxy anion, being twofold
degenerate and largely oxygen 2p-like (lone pair) with some hybridization with the
antibonding π*-orbital of the C-O group. We denote transition F as O1s→2e[π π π π*(C-O)]. For
multilayer methanol this signal is missing because the 2e orbital is filled . For pure HMP
(spectrum e) this signal is very weak, i.e. the 2e-orbital is completely filled. In other words,
the methoxy group of the pure donor material HMP is fully anionic. However, when the
orbital is not fully occupied the transition channel opens. This is obviously the case for the
HMP2-TCNQ1 complex (spectrum f) and - to a weaker extent - also for the other two
complexes (spectra g and h). The oscillator strength of transition F (essentially O1s→2p) is
very high because it is dipole-allowed and characterized by a large overlap of initial- and
final-state wavefunctions in the matrix element. This transition was also observed for the
surface methoxy species on Cu and Ni. The data of  reveal that the 2e-vacancy of this
species is quantitatively different on Ni(111) and Cu(111). For the latter case an effective
electron population of 3.6e instead of 4e has been derived .
The second signal G at about 536 eV is associated with the transition O1s→6a1[σ σ σ σ*(C-O)]. Its
intensity behaviour is similar to signal F. The methoxy σ*-resonances are observed also for
multilayer films of methanol . However, there is no π*-resonance for methanol . The
remaining signals H, I and K occur also for pure HMP and correspond to transitions into σ*
and π*-orbitals of the aromatic ring system. The bottom panel of Figs. 2 and 3 show the
spectra for the TMPx-TCNQycompound with 2:1 stoichiometry. This oxygen spectrum is the
one with the highest O 1s → 2e resonance intensity found in this study.
The transition F is thus a perfect indicator of the ionicity of the methoxy group in the donor
molecule. In Fig. 3 this signal increases strongly in the sequence of spectra h, g, f, i.e. for
increasing HMP content. On the other hand, for pure HMP (spectrum e) the signal is missing.
In the following section we will attribute the strong rise of signal F with increasing donor
content in the compound to an according increase of the vacancy in the HOMO of the
methoxy anion (2e orbital). In addition, the fit revealed small but significant shifts of the
resonance positions in Fig. 3 towards higher photon energies, opposite to the shifts in Fig. 2.
Figure 3: Same as Fig. 2 but for the K-edge NEXAFS spectra of oxygen. Bottom panel: Same
We now quantify the eye-catching result in the nitrogen and oxygen spectra (Figs. 2 and 3)
that the intensities of several resonances vary strongly with increasing HMP content. Fig. 4
summarizes the intensity variations (a, b) and the energy shifts (c, d) of the oxygen and
nitrogen resonances as function of the HMP/TCNQ ratio. The intensities have been
determined as ratios of the areas under the corresponding fit curves in Figs. 2 or 3, normalized
to the areas of peak A for nitrogen and peak I for oxygen (that are both not involved in the
charge transfer). The increase of the intensities of resonances 2e (F) and 6a1(G) in the oxygen
spectra (Fig. 4a) for the 2:1 compound counteracts the intensity decrease of resonances b3g
and au(C) and b1g and b2u (B) in the nitrogen spectrum (Fig. 4b). These intensity variations
with opposite behaviour for donor and acceptor give direct evidence of the participation of
specific orbitals in the charge-transfer process in these compounds. The strong variations are
the most direct indication of the intermolecular charge transfer. In particular the O 1s → 2e
transition directly mirrors the charge depopulation in the HOMO of the methoxy moiety.
Figure 4: NEXAFS resonance intensities in the oxygen (a) and nitrogen spectrum (b) and
energy shifts of the resonances in the oxygen (c) and nitrogen spectrum (d). Areas of the fitted
peaks B, C and D are normalized to area A for nitrogen and peaks F, G and H are normalized to
area I for oxygen. The energy shifts are referenced to the resonance positions in pure HMP and
A more indirect fingerprint is found in systematic shifts of the resonance positions as function
of stoichiometry. Fig. 4c and d shows the energy shifts of the resonance positions for oxygen
and nitrogen as determined from the fitted spectra. As reference we used the resonance
positions of the pure donor and acceptor (stoichiometry 1:0 and 0:1, respectively). The shifts
vary continuously with the HMP/TCNQ ratio and show an opposite sign for the donor and
acceptor moiety. Since there are several transitions hidden in peak H, we did not include it in
Fig.4c. The shifts are about a factor of 2 larger for the oxygen resonances than for the nitrogen
Let us recall the different mechanisms contributing to shifts in the NEXAFS features (see, e.g.
[8, 18]). The so-called valence shift reflects the redox state of the emitter atom: A change of
the redox state changes the screening effect of the valence electrons on the nucleus. In turn,
the binding energy of the electrons in the core level changes and the edge position appears
shifted. In addition, a chemical shift is induced by different ligands, analogous to the chemical
shift observed in XPS. Ligands with higher electronegativity reduce the effective charge
density in the region of the excited atom and thus cause a shift of the NEXAFS resonances
towards higher energies. Quantitatively, chemicals shifts are usually smaller in NEXAFS than
in XPS, because the XPS probes the N→N-1 transition of the system, whereas in NEXAFS a
transition to a neutral excited complex N→N* is observed, where both the core level and the
unoccupied final state undergo a chemical shift. The shifts of initial and final states may
partly cancel in the excitation energy.
The shift of all oxygen-related resonances towards higher photon energies (Fig. 4c) indicates
an increasing deficiency of valence charge leading to reduced screening of the ion core. Vice
versa, the shift of all nitrogen resonances to lower photon energies (Fig.4d) is indicative of an
increasing charge density in the valence region. It is interesting to note that the variations in
spectral weight of the resonances (Fig. 4a, b) and in the resonance positions (Fig. 4c, d) do not
show the same quantitative behaviour: The spectral weight shows a sudden change between
the 1:1 and 2:1 stoichiometry, whereas the shifts vary smoothly. This reflects the fact that the
occupation of the final states of the NEXAFS transitions are affected by the partial charge
transfer into specific orbitals, whereas the energetic positions vary because of the change in
net charge of the whole molecule.
Figure 5: Transition scheme describing the action of the charge transfer from the donor to the
acceptor. The NEXAFS spectra reveal 2e and 6a1 of the methoxy-group in TMP/HMP as
donating orbitals and b1g and b2u of the cyano-group of TCNQ as the accepting orbitals.
A surprising fact is that the charge deficiency in the donor (HMP) is highest for the compound
with maximum HMP content (2:1). The same is found for the TMPx-TCNQycompound with
stoichiometry 2:1 (see lowest panel in Fig. 3). In a simple model the opposite effect might be
expected. As the different stoichiometry leads to a different coordination and consequently to
a different structure, we conclude that the relative arrangement of HMP and TCNQ may play
an important role for the amount of charge transfer.
Fig. 5 summarizes the results in terms of a transition scheme including the charge transfer
scenario (schematic). The partial charge transfer denoted by the fat arrow explains the
intensity changes and energetic shifts of the observed NEXAFS resonances. Surprisingly, the
states 6a1 as well as b3g and aucontribute to the charge transfer process although they are
expected to be completely empty. This might be caused by the fact that NEXAFS does not
probe the pure ground state properties. During the excitation all states are shifted to lower
energies as a result of the core-hole attraction and therefore appear partially occupied.
5. Summary and Conclusion
Near-edge X-ray absorption fine structure (NEXAFS) spectroscopy was established as a tool
to study the orbital specific charge transfer in novel donor-acceptor compounds. The novel
functionalized pyrene-derivatives 4,5,9,10-tetramethoxypyrene (TMP, C20H18O4)
2,4,5,7,9,10-hexamethoxypyrene (HMP, C22H22O6) were employed as donors in complexes
with the classical strong acceptor 7,7,8,8-tetracyano-p-quinodimethane, (TCNQ, C12N4H4).
TMPx-TCNQy and HMPx-TCNQy crystals in stoichiometries x:y = 1:2, 1:1 and 2:1 were
grown by vapour diffusion of hexane into a dichloromethane solution of the components.
Oxygen and nitrogen K-edge NEXAFS spectra were exploited as local probes of the donor
and acceptor molecules, respectively. Oxygen is only contained in the methoxy-moiety of the
donor and nitrogen in the cyano-moiety of the acceptor. The orbital selectivity of the
NEXAFS resonances allows to precisely elucidate the participation of specific orbitals in the
The spectra revealed partial charge transfer from the methoxy-orbitals 2e (π*) and 6a1(σ*) of
the donor to the cyano-orbitals b3g and au(π*) and - to a weaker extent - to b1g and b2u (σ*) of
the acceptor. In particular, the occupation of 2e (being the HOMO of the methoxy-species
without charge transfer) reflects the anionic character of the methoxy-groups. Surprisingly,
the charge transfer is maximum for the compounds with highest HMP or TMP contents. As
additional indirect signature of a charge transfer, all spectral features of the donor and
acceptor are shifted to higher and lower photon energies, respectively.
Providing quantitative access to the relative occupation of specific orbitals, the NEXAFS
approach constitutes the most direct probe of the charge-transfer mechanism in organic salts
found so far. Although demonstrated for the specific example of pyrene-derived donors with
the classical acceptor TCNQ, the method is very versatile and can serve as routine probe for
novel CT-complexes on the basis of functionalized polycyclic aromatic hydrocarbons.
The project is funded through Transregio SFB TR 49 (Frankfurt, Mainz, Kaiserslautern),
Graduate School of Excellence MAINZ and Centre for Complex Materials (COMATT),
Mainz. We thank C. Felser, S. Naghavi (Inst. for Inorganic and Analytical Chemistry, Univ.
Mainz) and M. Huth, V. Solovyeva and M. Rudloff (Univ. of Frankfurt/Main) for fruitful
cooperation. We acknowledge the ANKA Angströmquelle Karlsruhe for the provision of
* corresponding author
 J. Wu, W. Pisula, and K. Müllen, Chem. Rev. 107 (2007) 718
 R. Rieger et al., Chem. Eur. J., 14 (2008) 6322-6325
 K. Medjanik et al., Phys. Chem. Chem. Phys., 12 (2010) 7184 – 7193
 D. Chercka, Diploma Thesis, 2010, Univ. Braunschweig
 E. Clar, W. Schmidt, Tetrahedron 32 (1976) 2563
 K. Medjanik et al., Phys. Rev. B 82 (2010) 245419
 7. T.-C. Tseng, C. Urban, Y. Wang, R. Otero, S. L. Tait, M. Alcami, D. Ecija, M. Trelka,
J. M. Gallego, N. Lin, M. Konuma, U. Starke, A. Nefedov, A. Langner, C. Wöll, M.A.
Herranz, F. Martin, N. Martin, K. Kern, and R. Miranda, Nat. Chem. 2 (2010) 374-379.
 J. Stöhr, NEXAFS spectroscopy, Springer Series in Surface Science 25 (1992)
 J. Fraxedas et al., Phys. Rev. B 68 (2003) 195115
 M. Sing et al., Phys. Rev. B 76 (2007) 245119
 11. R. Claessen, M. Sing, U. Schwingenschlögl, P. Blaha, M. Dressel, and C. S.
Jacobsen, Phys. Rev. Lett. 88 (2002) 096402.
 M. Sing, U. Schwingenschlögl, R. Claessen, P. Blaha, J. M. P. Carmelo, L. M. Martelo,
P. D. Sacramento, M. Dressel, and C. S. Jacobsen, Phys. Rev. B 68 (2003) 125111
 L. R. Melby et al., J. Am. Chem. Soc. 84 (1962) 3374-87.
 T. Amano, H. Kuroda, and H. Akamatu, Bulletin of the Chemical society of Japan, 44
 K. Amemiya et al., Phys. Rev. B 59 (1999) 2307
 S. Alibert-Fouet, I. Seguy, J. F. Bobo, P. Destruel, H. Bock, Chem. Eur. J. 13 (2007)
 Y. Yamashita, M. Tomura, J. Mater. Chem. 8 (1998) 1933.
 G. Bunker, Introduction to XAFS, Cambridge Univ. Press (2010)