Alkali Doping Leads to Charge-Transfer Salt Formation in a Two-Dimensional Metal–Organic Framework

Efficient charge transfer across metal–organic interfaces is a key physical process in modern organic electronics devices, and characterization of the energy level alignment at the interface is crucial to enable a rational device design. We show that the insertion of alkali atoms can significantly change the structure and electronic properties of a metal–organic interface. Coadsorption of tetracyanoquinodimethane (TCNQ) and potassium on a Ag(111) surface leads to the formation of a two-dimensional charge transfer salt, with properties quite different from those of the two-dimensional Ag adatom TCNQ metal–organic framework formed in the absence of K doping. We establish a highly accurate structural model by combination of quantitative X-ray standing wave measurements, scanning tunnelling microscopy, and density-functional theory (DFT) calculations. Full agreement between the experimental data and the computational prediction of the structure is only achieved by inclusion of a charge-transfer-scaled dispersion correction in the DFT, which correctly accounts for the effects of strong charge transfer on the atomic polarizability of potassium. The commensurate surface layer formed by TCNQ and K is dominated by strong charge transfer and ionic bonding and is accompanied by a structural and electronic decoupling from the underlying metal substrate. The consequence is a significant change in energy level alignment and work function compared to TCNQ on Ag(111). Possible implications of charge-transfer salt formation at metal–organic interfaces for organic thin-film devices are discussed.


Experimental Details, NIXSW data analysis and STM image simulation
The single crystal Ag(111) sample was cleaned in situ using cycles of sputtering with 1 keV Ar + ions for 30 minutes followed by annealing to ~500 °C for another 30 minutes. TCNQ powder (98% pure; Sigma Aldrich) was triply purified by thermal gradient sublimation and thoroughly degassed before being deposited onto the Ag(111) substrate from a Knudsen effusion cell at a nominal operating temperature of 120 °C. K was deposited by resistive heating of a SAES dispenser source. STM images were recorded in constant current mode using electrochemically etched polycrystalline tungsten tips. All STM images were plane corrected and flattened using the open source image-processing software Gwyddion 1 .
Synchrotron radiation X-ray photoelectron spectroscopy (SXPS) characterisation and normal incidence X-ray standing wave (NIXSW) measurements were performed in the UHV end-station installed on beamline I09 of the Diamond Light Source. This beamline is equipped with both a crystal monochromator to provide the 'hard' X-rays needed to perform the NIXSW experiments and a grazing incidence plane grating monochromator delivering soft X-rays to perform highresolution SXPS. Both the SXP spectra and the higher photon energy XP spectra used in the NIXSW measurements were recorded using a VG Scienta EW4000 HAXPES hemispherical electron analyser mounted at 90° to the incident photon beam, while sweeping the photon energy through the (111) Bragg condition at near-normal incidence to the surface. The high-resolution SXP spectra and the XP spectra recorded in NIXSW experiments were fitted using the CasaXPS software package. Table S1 shows a comparison of the photoelectron core level binding energies derived from the SXP spectra of the K2TCNQ overlayer and the Ag-TCNQ surface layer. The NIXSW technique 3 exploits the X-ray standing wave formed by the interference of incident and scattered X-rays at a Bragg condition. The strong backscattering out of the crystal leads to a finite extinction depth of the X-rays, resulting in high reflectivity over a finite range of incident X-ray energies. Within this range the standing wave, which has a periodicity in intensity equal to that of the scattering planes, shifts in phase relative to the scattering planes in a systematic fashion. Monitoring the X-ray absorption at atoms immersed in the standing wave (within or above the crystal) through this photon energy range thus provides a profile characteristic of the height of the absorbing atoms above the scattering planes. The use of photoemission to monitor the absorption allows one to exploit the element and chemical-state-specific photoelectron binding energies in order to extract structural data from each locally inequivalent atom in the overlayer. The measured NIXSW absorption profiles can by uniquely fitted by two structural parameters, the coherent fraction, f, and the coherent position, p. The coherent fraction is commonly regarded as an order parameter, with a value of unity for a single adsorption geometry that is perfectly (statically and dynamically)  Figure S1 shows the raw NIXSW data obtained from the Ag(111)-K2TCNQ surface, while    Table   S2. A similar effect could account for the f value of the N atoms, which is the same as that of the C atoms ('CN') bonded to N; notice that the atomic height difference between the pure TCNQ and coadsorption phases are largest for the CN and N atoms, so some admixture of the two phases would lower these measured f values the most.
The very low f value for N in the Ag-TCNQ phase has been shown to be due to two distinctly different N heights in a twisted molecule, arising from the presence of the Ag adatoms in the overlayer structure; 2 in the K2TCNQ phase this parameter is not so anomalously low. The high f value for the main K peak, and the much lower value for the minor K component, are fully consistent with the suggestion in the main text that the main peak corresponds to K atoms at a single well-defined height in the 2D-MOF, while the minor K component is related to excess K atoms that may lie above the 2D-MOF (possibly at a height 2.35 Å higher than the value reported in the table, with a n value of 2 rather the value of 1 used in the  Table S3. These calculations take no account of the dynamic and static disorder of the experiment (including possible partial coexistence of two surface phases) and are thus all higher than the experimental values.
The predicted values of the coherent positions are given in the main manuscript in    Fig. S2 overlaid on a section of the experimental image of Fig. 1(a). Simulated images using this technique invariably show higher spatial resolution than is found in experimental STM images, even when formed using very low isosurface values -

Computational Details of Density Functional Theory calculations
The theoretical analysis has been performed at the density functional theory (DFT) level using the Fritz-Haber institute ab initio molecular simulations package (FHIaims) 8  Hirshfeld charge partitioning calculations were performed using the "tight" basis set functions however, the numerical integration grid was replaced with the "really tight" settings. This was necessary to reduce integration errors in the calculation of atomic net charges in both methods.

Details of dispersion parameter rescaling for potassium
Within this work we employed two distinct dispersion correction schemes, vdW surf and MBD. Both of these schemes use atomic reference values for atomic polarisability, C6 coefficient and van der Waals (vdW) radius for neutral free atoms as input data. These tabulated parameters are then rescaled based on the ratio between the effective atoms-in-molecules (AIM) volume and the free atom volume: This scheme is known to underestimate the effect of strong charge transfer in polar chemical bonding situations. To better describe long-range dispersion interactions involving K + , we replace the free atom reference polarizability of neutral K with the one for K + and we add a scaling pre-factor to change the reference volume:

Calculation of energy level diagrams of K2TCNQ and Ag-TCNQ on Ag(111)
Molecular orbital projected density of states (MODOS) provide a method of projecting gas phase molecular orbitals onto the electronic structure of a metalorganic interface. The density of states (DOS) of a quantum mechanical system as a function of energy is defined as, where, is an energy eigenstate of the system and is the corresponding energy eigenvalue. The utility of such a method allows a projection onto a complete orthogonal basis set { ( )} without loss of information.
If the single state ( ) belongs to one of the molecular orbitals (MOs) of a molecule in the gas phase, this gives an MO-projected DOS or MODOS.
The results are shown in figure 6 of the main manuscript.    Table S8 provides a breakdown of work function values for the total system, the pristine metal surface and the difference between the adsorbate-substrate system and the pristine substrate work function, respectively. We can decompose the change in work function for both systems into contributions from the electrostatic dipole moment of the overlayer (Δ mol ) and from the bond between the metal and the overlayer (ΔE bond ). We find that the bond contribution is slightly larger for the Ag-TCNQ case, but the dominant difference between the two is the large electrostatic potential drop introduced by the dipole density of the K2TCNQ salt layer.

Density difference plots of K2TCNQ and Ag-TCNQ on Ag(111)
Density difference contour plots (Δ ) were calculated using equation S6 for