2.1 Ultrafast Dynamics in Solids and at Interfaces
The study of the non-equilibrium, ultrafast dynamics in solids and at interfaces gives fundamental insights into elementary processes like electron transfer and localization, electron-phonon coupling and energy dissipation in complex systems. Real-time studies of such processes are performed by several groups in the department using various ul-trafast laser techniques.
2.1.1 Dynamics of Interfacial Electron Transfer
The transfer of charge and/or energy across the interface between two materials is the key for many functionalities like light harvesting in solar cells, light emitting devices or electrochemical processes. A detailed understanding of electron transfer dynamics and competing relaxation channels like electron scattering or localization is thus highly relevant. Heterogeneous electron transfer is determined by the coupling matrix element between the donor and the acceptor state (e.g. in the adsorbate and substrate) and occurs on characteristic (typically femto- to picosecond) time scales. This electronic coupling can be significantly modified by the response of the molecular surrounding to a photoin-jected electron if its lifetime is comparable to the timescale of polarization and reorganization of the adsorbate. Elementary processes like polaron formation and electron solvation can therefore directly compete with the electron transfer, as they modify the electronic coupling across the interface.
The dynamics of charge transfer, electron localization, and relaxation in molecular ad-layers on metal and semiconductor surfaces have been investigated at the FU Berlin and by the group of Julia Stähler using femtosecond time- and angle-resolved two-photon photoemission (2PPE). This technique enables the direct observation of excited state population dynamics and measurement of absolute binding energies as a function of parallel momentum, providing information about the dispersion of electronic states.
(i) Ultrafast charge transfer from alkali-doped ice layers on Cu(111): In the last few years amorphous and crystalline ice adsorbed on single crystal metal surfaces has been used as a model system to systematically investigate the dynamics of electron transfer and solvation processes at interfaces. In these experiments, photoinjection of electrons from the metal into the ice conduction band is followed by ultrafast localization and solvation of the excess electrons. The subsequent energetic stabilization of these solvated electrons due to nuclear rearrangements of the polar molecular environment is ac-companied by an increasing degree of localization of the electron charge density. For amorphous D2O layers the population of solvated electrons decays non-exponentially on timescales of several 100 fs .
Fig. 1: Electron solvation and relaxation dynamics at alkali/D2O/Cu(111) interfaces probed by time-resolved two-photon-photoelectron spectroscopy . (a) For pure amorphous multilayers electron transfer back to the Cu substrate occurs on a timescale of a few 100 fs. (b) Co-adsorption of small amounts of sodium on top of the ice film leads to the formation of long-lived species. (c) Schematic illustration of different electron species in the D2O film (left) and at Na+/D2O complexes at the ice/vacuum interface (right). (d) Buildup of the solvation shell around individual Na adsorbates at low water coverage (> 6 molecules per Na atom).
Recently the influence of alkali ions on the electron solvation dynamics has been stud-ied for two types surface preparations: (i) sub-monolayer coverages of alkali atoms on top of preadsorbed amorphous D2O multilayers (i.e. adsorption at the ice-vacuum inter-face) and (ii) alkali adsorption at the metal substrate and subsequent buildup of a solvation shell by exposure to small amounts of water. Experiments were performed for Cs, K and Na as a function of coverage. Fig. 1 shows that the alkali adsorption following preparation (i) (0.08 ML Na on top of 5 BL D2O) leads to pronounced changes in the observed electron dynamics compared to bare amorphous ice. A new feature resulting from long-lived electrons appears in time-resolved 2PPE spectroscopy which exhibits much longer lifetimes (of the order of 1 - 5 ps) and a different (faster) energetic stabilization compared to the solvated electrons in pure D2O ice (see Fig. 1 a and b). This is attributed to the formation of a transient electron-sodium-ion-complex which is located at the ice/vacuum-interface as illustrated in Fig. 1 c). The location of the excess electron at the ice-vacuum interface decouples the wave function from the metal leading to prolonged lifetimes. This interpretation is further corroborated by coverage dependent measurements and by overlayer experiments.
The second type of experiments (alkali pre-covered Cu(111) exposed to small amounts of D2O) makes it possible to study the buildup of the solvation shell around the alkali ions and its influence on the dynamics of unoccupied electronic states. Above a certain number of D2O molecules per alkali atom a new feature is observed in 2PPE spectros-copy in addition to the well-known unoccupied alkali resonances on Cu(111). This feature is attributed to an excess electron bound to an alkali-(D2O)n cluster (see Fig. 1 d). For Cs/Cu(111) such trapped electrons are observed above a minimum water coverage equivalent to n = 2 - 3 D2O molecules per Cs atom, whereas for K/D2O and Na/D2O clusters the coverage to stabilize an excess charge corresponds to n = 5 - 6 and n = 6 - 7 water molecules per alkali atom, respectively. Furthermore, the energetic stabilization occurs much faster compared to solvated electrons in amorphous ice indicating the different electronic nature of the solvated electrons. Currently theoretical investigations are underway to gain insights into the structure of such alkali-water complexes.
(ii) Electron transfer dynamics at hybrid inorganic-organic interfaces: Within the re-cently established collaborative research center Sfb 951, the group of Julia Stähler investigates the electronic structure and charge carrier dynamics at interfaces of hybrid inorganic-organic systems (HIOS) and has carried out 2PPE experiments of pyridine/ZnO(10-10), which serves as a benchmark system for theory. The experiment reveals a finite density of states at the Fermi level (i.e. the formation of a metallic interface) as well as an unoccupied interface state 1.7 eV above EF in accordance with preliminary DFT calculations of Patrick Rinke in the theory department. In addition, time-resolved 2PPE measurements suggest a delayed charge injection into the ZnO conduction band on picoseconds timescales via this interface state and an upward band-bending at the pyridine/ZnO(10-10) interface of ~0.5 eV.
Future studies will focus on small organic dyes like oligo-phenyls or fluorenes to inves-tigate the ultrafast dynamics of carriers and excitons at the hybrid interface. Complementary to these 2PPE studies, the dynamics at HIOS interfaces will be investigated using femtosecond time-resolved optical spectroscopies like transient absorption and time-resolved electronic sum frequency generation (TRE-SFG) providing information on the interfacial electronic structure and carrier dynamics.
||J. Stähler, U. Bovensiepen, and M. Wolf: In: Dynamics at Solid State Surfaces and Interfaces, Vol. 1, (Eds.)
U. Bovensiepen, H. Petek, and M. Wolf. Wiley-VCH, Berlin (2010), p 359-379.|
||M. Meyer, M. Bertin, U. Bovensiepen, D. Wegkamp, M. Krenz and M. Wolf, J. Phys. Chem. C 115, 204 (2011).|
2.1.2 Transient Electronic Structure of Correlated Materials
Strongly correlated electron materials exhibit exotic electronic and magnetic properties, characterized by broken-symmetry ground states such as metal-to-insulator instabilities, unconventional superconductivity, and various cooperative ordering phenomena. One of the major challenges in this field is to understand the ground and excited state properties on a microscopic level and disentangle the competing interactions and correlations of charge, spin, orbital and lattice degrees of freedom, which are acting on multiple length, energy and time scales. Femtosecond time-resolved spectroscopy techniques are a powerful tool for the investigation of elementary scattering processes in such complex materials. For example, in a material with metal-to-insulator transition, optical excitation can result in a transient closing of band gaps where the timescale of the gap closing is char-acteristic of electronically or vibrationally driven processes.
Time-and Angle-Resolved Photoemission Spectroscopy (tr-ARPES) extends the benefits of momentum-resolved electron spectroscopy into the time domain and provides direct access to elementary electron scattering processes by optically exciting the electronic system and subsequently probing the evolution of the transient electronic structure by photoemission.
We have applied tr-ARPES to the study of the ultrafast electron dynamics of the unconventional high-Tc superconductor Bi2Sr2CaCu2O8+d far below its critical temperature , see Fig. 2. The major finding is that the density of non-equilibrium quasi-particles created by photo-induced breaking of Cooper pairs is momentum-dependent and related to the size of the superconducting gap, while their recombination rate is independent of momentum or excitation density. This shows the transient stabilization of quasi-particles off the node due to phase space restrictions, which are caused by energy and momentum conservation in a d wave superconductor in a boson-bottleneck regime. We emphasize that our studies of weakly perturbative excitations of cuprate compounds without destroying the superconducting state are enabled by an excellent signal-to-noise ratio, only achievable with a source operating at several 100 kHz repetition rate.
Fig. 2: tr-ARPES of Bi2Sr2CaCu2O8+d . (a) Sketch of the normal state Fermi surface. (b) ARPES inten-sity, measured along the red arc in (a). (c) tr-ARPES spectra at selected pump-probe delays. The relative depletion of the superconducting peak and increase of spectral weight above EF are shadowed in yellow and red.
As a second example our tr-ARPES study of the Fe-pnictide parent compound EuFe2As2, serving as a model system for unconventional iron-based superconductors, reveals a momentum dependent carrier dynamics around the hole pocket in the center of the Brillouin zone . The different dynamics of electrons and holes point to interband scattering processes in addition to intra-band relaxation. Furthermore, three coherent modes are observed in transient modulations of the electron distribution. One important result is that a quantitative analysis of the relaxation processes provides an upper bound for the electron-phonon coupling constant of l < 0.5. This result indicates a minor influence of electron-phonon coupling for the pairing mechanism leading to superconductivity in the Fe-pnictides.
On the instrumental side, considerable efforts are currently invested in designing and setting up an advanced tr-ARPES experiment. The groups of Ralph Ernstorfer and Patrick Kirchmann are collaborating intensely to construct a novel femtosecond high harmonic generation (HHG) source with an exceptionally high repetition rate of several 100 kHz. This system is based on a high-power fiber-laser and provides sufficient pulse energy for non-collinear optical parametric chirped-pulse amplification in order to subsequently drive efficient HHG in rare gas targets. This unique XUV source will be combined with a state-of-the-art ARPES ultrahigh vacuum system, featuring an imaging-type hemispherical electron analyzer and an ultra-stable cryogenic 6-axis manipula-tor. The benefits of our approach for HHG-based tr-ARPES are multifold: the virtually unlimited access to the Brillouin zone enables addressing the full electronic structure of many relevant materials with tr-ARPES (see Fig. 3). The simultaneously improved time resolution facilitates detection of even faster phenomena such as high-frequency coherent modes. At the same time, we maintain the superb statistics only achievable with high repetition rate sources.
Fig. 3: Fermi surface of CeTe3 measured at 25 K using synchrotron radiation at h = 55 eV showing the effect of different matrix elements in the 1st and 2nd Brillouin zone (BZ) (see PRL 93, 126405) Present tr-ARPES setups with h = 6 eV are limited to the center of the 1st Brillouin zone (BZ), indicated by red arcs for 20 , 40 and 60 emission angle. In contrast, fs-XUV light provides full access to the 1st and 2nd BZ where potentially different dynamics can be observed.
In short, our efforts aim at significantly furthering the applicability of tr-ARPES in or-der to contribute to our basic understanding of solid state physics and help answer noto-rious materials science questions. Along these lines, Patrick Kirchmanns group will continue our fruitful studies of unconventional superconductors [1, 2], prototypical charge density wave systems of the rare-earth tri-telluride family and metallic quantum well systems  and extend our collaborations and research projects in the future. In particular, tr-ARPS studies of topics such as self-assembling metallic nano-wires, topological insulator compounds, conventional metallic super-conductors, and d- and f-shell transition metal oxide surfaces will become feasible with our new experimental infrastructure.
||R. Cortés, L. Rettig, Y. Yoshida, H. Eisaki, M. Wolf, and U. Bovensiepen, Phys. Rev. Lett. ( in press).|
||L. Rettig, R. Cortés et al., arXiv:1008.1561.|
||P. S. Kirchmann, et.al., Nature Physics 6 (2010) 782.|
2.1.3 Interactions of Electrons and Phonons in Photoexcited Solids
The optical and electronic properties of solids are governed by the interplay of electrons and ion cores. The electronic and phononic band structure depend on the symmetry of the crystal potential, which also determines the Brillouin zone boundaries. A change of lattice symmetry can therefore change the conductivity if a band gap is formed at the Fermi level (EF). Such Peierls-type distortions of the lattice go along with a redistribution of the electron density following the reduced symmetry, lowering the electronic energy and resulting in semiconducting or insulating behavior. On top of the electron-lattice interaction, however, also electronic correlation effects can significantly affect the conductivity of a material. When the Coulomb repulsion between the electrons is sufficiently high, they can lower their energy by (Mott) localization also leading to insu-lating behavior and the formation of a band gap at EF.
Vanadiumdioxide (VO2) undergoes an insulator-to-metal transition as a function of temperature that goes along with a structural transition from a low-symmetry mono-clinic (M1) to the higher symmetry rutile (R) phase. As known from previous time-resolved studies, the transition can be induced by photoexcitation, however, due to the limited time resolution of the lattice probing techniques (X-ray and electron diffraction), the onset of the structural transition, which occurs along with the ultrafast excitation of the electronic system (femtosecond timescales), could not be investigated so far.
Making use of the coherent phonon response of the material, the group of Julia Stähler has developed an all-optical technique that is able to probe changes to the lattice potential symmetry on femtosecond timescales. The underlying principle is illustrated by the respective phonon band structure shown in insets of Fig. 4 a) and b). The symmetry of a crystal (exemplified by the different Brioullin zone boundaries at Z and Z' ) determines the phononic band structure and thereby also the number of Raman active phonons at the G point. A photoinduced symmetry change should therefore lead to a change in the optical phonon spectrum, which can thus be utilized to probe the crystal lattice symmetry.
Fig. 4: Transient reflectivity of VO2 as a function of probe wavelength below (a) and above (b) the threshold of the photoincduced phase transition. Below threshold, signals at longer wavelengths are dominated by oscillations superimposed with a negative transient. Above threshold, no oscillations are observed, showing that the phonon modes are lost over the entire visible spectrum. Panel (c) and (d) show the evolution of the reflectivity at a probe wavelength of 525 nm for excitation with a single pulse and pulse pair with a delay of one half or full oscillation period. Above threshold, the phonon mode cannot be re-excited with the second pulse.
Indeed, weak excitation below the threshold fluence of the phase transition leads to pro-nounced oscillations in the transient optical reflectivity due to the coherently excited optical phonons (Fig. 4a), as measured over the whole visible spectrum using a white light supercontinuum probe pulse . These oscillations vanish when the VO2 is excited above threshold (Fig. 4b). Whether this finding must be attributed to dephasing or to a photoinduced increase in symmetry of the sample is analyzed in a three-pulse experiment (see Fig. 4 c and d): Thereby, a first pump pulse P1 creates non-equilibrium conditions in the material and a second pump pulse P2 re-excites the sample, which is subsequently probed with hprobe. For excitation below threshold, i.e. if the phase transition is not driven by P1, the second pump can, depending on its relative phase to P1 (p and 2p) , quench or enhance the amplitude of the coherent oscillations, respectively, because the symmetry of the equilibrium phase is maintained (Fig. 4c). In contrast, for excitation above threshold, P2 cannot re-excite coherent lattice vibrations, independent of its relative phase: The strong photoexcitation has thus changed the symmetry of the lattice on a sub-phonon period timescale, at the onset of the photoinduced phase.
In future studies, the group is heading towards charge-injection into thin layers of strongly correlated materials followed by the collapse of the Mott-Hubbard band gap. The electronic structure change at in situ prepared interfaces will be probed by fs time- and angle-resolved two-photon photoemission (2PPE) spectroscopy. Also interface-sensitive non-linear optical spectroscopy will be employed using a 10 fs white light super continuum pulse as a probe .
Furthermore, tr-ARPES experiments have been performed at the FU Berlin on the pro-totypical charge density wave (CDW) systems of the rare-earth tri-telluride family . There, the dynamics of the photoinduced transition is observed via closing of the electronic band gap at EF and is governed by excitation of the so-called CDW amplitude mode. The size of the CDW gap is directly related to the amplitude of this mode. Recently, coherent control of the amplitude mode and gap size in TbTe3 could be demonstrated in a two pulse excitation scheme.
||D. Wegkamp, et al, Appl. Phys. Lett. (in press). ( in press).|
||F. Schmitt et.al., New J. Phys. 13 (2011) 063022.|
2.1.4 THz Spectroscopy of Low-energy Excitations
Many fundamental elementary excitations in nature have transition energies of the order of 10 meV. Examples are quasi-free electrons in solids, Cooper pairs in superconductors, and excitons in semiconductors. Electromagnetic pulses with frequencies in the terahertz (THz) range have been proven to be efficient probes of such excitations thanks to their low photon energy (4.1 meV at 1 THz), their large bandwidth (typically more than one octave), and the possibility to detect the transient electric field (rather than intensity). In addition, the short pulse duration (typically less than 1 ps) makes it possible to study ultrafast processes.
The group of Tobias Kampfrath employs time-resolved THz transmission spectroscopy where, e. g., charge-carrier dynamics in a thin sample is triggered by the absorption of a femtosecond laser pulse and, after some delay, probed by a broadband THz pulse. Besides using such well-established techniques, the group aims at extending THz spectroscopy towards the active control of low-energy excitations, like magnetic interactions, or low-energy excitations of interfaces of solids. Currently the following lines of research are pursued.
Fig. 5: Left: Schematic of THz electron spin resonance. An intense magnetic THz pulse resonantly drives a spin wave in antiferromagnetic NiO. The instantaneous sample magnetization is probed by a time-delayed near-infrared (NIR) pulse. Right: Coherent control of a 1-THz precession in NiO using different pulse sequences (one-pulse excitation; amplification or quenching by a second pump pulse arriving 6 or 6.5 precession cycles later, respectively).
Magnetic coupling: In optics, the interaction between THz pulse and sample is usually mediated by the electric field, whereas the usually weaker interaction with the magnetic field is not exploited. Magnetic coupling can be enhanced by making use of spin reso-nances or by means of tailored dielectric structures. The first approach has been recently demonstrated as schematically shown in Fig. 5 (left). Using the magnetic field of an intense THz pulse, a spin wave in the antiferromagnet NiO has been launched (see Fig. 5 (right)). A second pulse arriving 6 or 6.5 spin precession cycles later, respectively, was used to instantaneously amplify or stop the magnon . In a first experiment concerning the second scheme, the light-induced magnetic moment of a metallic ring was measured with a diameter smaller than the wavelength of light (in this case 1.5m) .
Interface sensitivity: So far, THz pulses have exclusively addressed the bulk properties of the samples (rather than their interfaces). The group plans to investigate the potential of THz spectroscopy in order to study low-energy excitations that are localized at interfaces. As the expected signals are small, significant effort has been invested increasing the signal-to-noise ratio. In a first experiment, ultrashort current pulses have been detected flowing along the interface of a ferromagnetic Fe film and a nonmagnetic cap layer. The current pulse was launched by excitation with a femtosecond laser pulse and detected by measuring the THz pulse arising from the current. The emitted THz wave-form was found to depend very sensitive to the cap material, demonstrating its interface sensitivity. This information might be highly relevant for spintronic devices as giant magnetoresistance relies on interface properties.
Nonlinear THz spectroscopy: Most THz studies to date have used THz pulses as a probe of low-energy excitations. THz pulses will be used to control low-energy excitations. For example, resonant pumping of infrared-active phonon modes provides a means of excitation that is fundamentally different than creating hot carrier distribution by absorption of visible photons at optical frequencies. The group has recently moved to the new lab featuring a new, cryogenically high-power laser system, which is a prerequisite for the generation of intense THz pulses.
Terahertz photonics: In contrast to optics in the visible spectral range, there is a lack of photonic devices that can be used to manipulate THz radiation. The arsenal of conventional elements (such as lenses and mirrors) will be extended to devices for filtering and frequency conversion. To realize a test bed for such devices, transient metallic structures will be used that are imprinted into a semiconducting substrate by means of a femtosecond laser beam with a shaped cross section. In particular, antenna-like structures will be explored  in order to locally enhance an incident THz pulse, which is interesting for nonlinear THz optics.
||T. Kampfrath, A. Sell, G. Klatt, A. Pashkin, S. Mährlein, T. Dekorsy, M. Wolf, M. Fiebig, A. Leitens-torfer, R. Huber, Nature Photonics 5, 31 (2011).|
||M. Burresi, T. Kampfrath, D. van Oosten, J. C. Prangsma, B.S. Song, S. Noda, L. Kuipers, Phys. Rev. Lett. 105, 123901 (2010).|
||I. Sersic, C. Tuambilangana, T. Kampfrath, A. F. Koenderink, Phys. Rev. B 83, 245102 (2011).|
2.1.5 Ultrafast Transport of Spin-polarized Hot Carriers
The ultrafast spin dynamics induced by transport of spin polarized carriers has attracted considerable interest over last decade. It is motivated by the fundamental interest in magnetic excitations and applications like spintronics and data storage. To achieve a microscopic understanding of the underlying elementary processes that typically occur on femtosecond time scales, the group of Alexey Melnikov has developed a time domain approach that probes the spin dynamics induced by hot carriers (HC) and demonstrated spin polarized HC transport through an epitaxial Au/Fe/MgO(001) structure .
Following a time-of-flight approach, which is realized in a back-pump/front-probe con-figuration (see Fig. 6 a), it has been shown that HC induced in Fe by the pump laser pulse can form a nearly ballistic spin current in Au (Fig. 6 b). Optical second harmonic (SH) is generated at the Au surface by the probe pulse and monitors the transient HC density in the subsurface layer as well as the transient surface spin polarization (SP) induced by HC. There are two contributions to the SH electric field, which are even and odd upon the reversal of SP, respectively (Fig. 6 c, d).
The HC pulse consists of steep leading part formed predominantly by ballistic HC propagating though the Au layer with the time-of-flight defined by the Fermi velocity and shallow part formed by HC propagating predominantly in the diffusive regime. The ballistic contribution is dominated by spin-up holes excited to final energies close to the Fermi level (see Fig. 6 b) and consequently having larger lifetimes, while the diffusive one is dominated by spin-up electrons excited to relatively high energies. Owing to that, the leading part of HC packet brings a negative SP to the Au surface providing a negative spike of SH magnetic contrast (Fig. 6 d). After reflection/scattering of these ballistic HC from the Au surface the trial part of HC pulse brings the positive SP leading to the sign change of . The transient SP decays at 1 ps time scale (Fig. 6 d), which is faster than the decay of HC density (Fig. 6 c). This timescale is attributed to the electron-electron spin-flip scattering of HC in Au and corresponds to the spin-flip probability of 0.1 in every single scattering event.
Fig. 6: (a) The scheme of our experimental approach. (b) Calculated density of HC excited in Fe vs. their energy with respect to the Fermi level: solid (dotted) curves correspond to carriers with negative (positive) spin polarization. Dashed curve reproduces the hot carrier lifetime according to Phys. Rev. B 58, 10948 (1998) (c, d) Pump-induced variation of SH electronic (even) contribution E (c) and transient SH magnetic contrast ~ SP calculated as a ratio of odd to even SH contributions (d) measured at the Au surface of Au/Fe/MgO(001) structure with indicated thicknesses of Fe and Au layers for the pump fluence of 1 mJ/cm2.
In the future this approach will be extended to three-layer structures involving two ferromagnetic layers separated by a paramagnetic spacer, allowing parallel and anti-parallel orientation of the magnetization in ferromagnetic layers. The first candidate is MgO/Fe/Au/Fe/MgO(001) with different thicknesses of Fe layers, which will allow to excite the spin dynamics in Fe film by spin-polarized HC with the SP oriented parallel or anti-parallel to the magnetization of the probed film.
||A. Melnikov, I. Razdolski, T. Wehling, E. Papaioannou, V. Roddatis, P. Fumagalli, A. Aktsipetrov, A. Lichtenstein, and U. Bovensiepen, Phys. Rev. Lett. 107, 076601 (2011).|
2.2 Molecular Processes at Surfaces
Studies of molecular processes at surfaces are performed by several groups in the department which employ complementary techniques with high spatial resolution as well as with chemical sensitivity using vibrational spectroscopy. In these studies, surface reactions and molecular rearrangements are stimulated by thermal activation and excitations by light or charge transfer.
2.2.1 Nanoscale Science with Functional Molecules
The research activities in the group of Leonhard Grill focus on the investigation of inorganic and organic matter, in particular functional molecules, on surfaces by low temperature scanning tunnelling microscopy (STM), preferentially at 5 K. This method makes it possible to image single molecules with a spatial resolution below 100 pm and is especially used for molecular manipulation by the Nanoscale Science group: Lateral or vertical displacement of single molecules or chemical reactions are induced with the STM tip, taking advantage of chemical forces, the tunnelling electrons or the electric field in the junction.
While the manipulation of atoms and molecules by STM is a well-known technique, the manipulation of inorganic crystallites has hardly been attempted in the past. By using an ultrathin crystalline film of NaCl on a Cu(111) surface, the research group could show that such a nanostructure can react in various ways on the stimulus of the STM tip . Depending on the crystallite size, NaCl nanostructures can be moved laterally or, if the island is large and the interaction with the metal surface is strong, it can be cut. Additionally, a cracking process can be induced for crystallites with a cantilever shape. Furthermore, an elastic regime of NaCl crystallites was found, where Hookes law is valid as proven by theoretical simulations.
Beside the fundamental understanding of physical and chemical processes at the atomic scale, the research group is interested in the formation of molecular assemblies that are stabilized by a well-defined intermolecular interaction. Depending on the chemical structure of the molecules, rather weak van der Waals interaction, hydrogen bonds or strong interactions such as metal-ligand or covalent carbon-carbon bonds are achieved. In the latter case, polymerization occurs directly on the surface, leading to macromolecules, which are more promising than supramolecular structures, in view of stability and potential charge transport.
Fig. 7: On-surface polymerization of nano-trains on a Cu(111) surface . (a) shows the scheme of connecting individual molecules in a linear fashion. The chemical structure of a molecular building block is presented in (b) with an STM image in (c). After connection by metal ligand bonds, chains of molecules are visible in the STM image (d).
The group has recently attempted to connect so-called wagon molecules, molecular building blocks with carborane side groups that should act as wheels and roll on the surface during diffusion (see Fig.7 a-c). While this rolling motion could not be proven, due to a rather strong interaction with the substrate, the molecular building blocks could be connected successfully to so-called nano-trains (Fig. 7 d) . A detailed analysis of the formed structures revealed that metal-ligand bonds stabilize these structures, including a copper atom from the substrate, which is even visible in the STM images (indicated by an arrow in the inset of Fig. 7 d).
An important class of molecules in terms of functionality are molecular switches that exist in at least two stable states. Recently, the group has also found evidence for an inverted thermal switching behaviour in imine-based molecules . As typically the trans state is energetically favoured, such molecules relax from cis to trans upon heating. On a gold surface however, the inverse behaviour is found, probably induced by the interaction with the substrate, leading to an increasing number of cis isomers after heating procedures.
||C. Bombis, et al., Phys. Rev. Lett. 104, 185502 (2010).|
||C.J. Villagómez, T. Sasaki, J. M. Tour, and L. Grill, J. Am. Chem. Soc. 132, 16848 (2010)..|
||J. Mielke, et al., ACS Nano 5, 2090-2097 (2011).|
2.2.2 Tip-enhanced Raman Spectroscopy
Tip-enhanced Raman spectroscopy (TERS) represents a promising tool to investigate interfaces topographically and with chemical and molecular sensitivity on a nanometer scale. The TERS approach bears a great potential for local identification and characterization of adsorbates by optical vibrational spectroscopy with high sensitivity and nanometer resolution.
The group of Bruno Pettinger has implemented TERS in UHV by employing a unique concept: (i) an adjustable high numerical aperture parabolic mirror is placed in between the STM scanner and the sample and (ii) all other necessary optics are mounted on a common platform together with the scanning tunnelling microscope (STM). In 2007 first very promising results were obtained, however, the instrument developed permitted only the proof of principle. More recently, the instrument was improved along various lines in order to make UHV-TERS more easily applicable and complementary for advanced UHV studies. This concerns, for example, the optical alignment for which piezo-driven mirrors have been installed permitting a re-alignment of the optical path without opening the UHV chamber. In addition, a preparation chamber has been added, which enables the sputtering and annealing of a single crystal sample, the evaporation of molecules and the transfer to the TERS unit in UHV for vibrational investigations.
Fig. 8: Time dependence of TERS on a C60 island at Au(111) at RT. Over 100 original spectra were recorded subsequently with 0.5 s acquisition time and grouped into nine average spectra, separated in time by 5s; Color code: from red to blue. Two bands marked by x represent artifacts. Laser power: 100 mW. Exciting line: 632.8 nm. Insert: STM image of a section of a C60 island.
Currently, new experiments are in progress. These include investigations on C60 molecules in form of well ordered islands deposited at room temperature on Au(111) samples (see Fig. 8). First results on such samples yield TER spectra with high signal-to-noise ratio that show much more Raman lines (>25) than allowed for isolated C60 molecules (where only 10 modes are Raman active). This is attributed to the symmetry reduction caused by adsorption and intermolecular interactions, which may lift the degeneracy. Thus four former IR-active modes as well as some other silent modes (among 32) may become not only Raman active but also relatively intense under TERS conditions. Most strikingly is a significant time dependence of the C60 TER spectra, not so much in observed vibrational frequencies, but in the relative band intensities. This could point to structural and chemical changes within the adsorbed C60 adlayer. These changes in the C60 TER spectra make this system an interesting test case for molecule-substrate and intermolecular interactions.
Sub-monolayers of dye-molecules (e.g. isotope substituted two-analyte systems) will be investigated next, in order to determine the spectroscopic properties of individual adsorbates. The measurements may be extended also toward optically non-resonant molecules adsorbed at single crystalline surfaces as well as individual nanoclusters.
Furthermore, we plan to study in close cooperation with the department of Chemical Physics supported nanoparticles on thin oxide films by TERS. For the preparation of such more samples a new preparation chamber has been constructed and will soon be implemented in the UHV system.
2.2.3 Photoinduced Surface Reactions and Vibrational Spectroscopy
The group headed by Christian Frischkorn has studied photoinduced chemical processes at surface in various directions over the last years. Besides femtosecond laser-induced surface reactions like associative desorption reactions of diatomic molecules (e.g. H+HH2; C+OCO), electron solvation dynamics in D2O ice on Ru(001) have been investigated using vibrational spectroscopy based on sum-frequency generation (SFG). In particular, the group performed experiments on the vibrational response of the polar D2O solvent molecules to excess electrons injected after UV laser excitation. As a result, in crystalline D2O ice layers a giant SF signal enhancement by 3 to 4 orders of magnitude was observed (see Fig. 9).
The explanation for this phenomenon is based on ferroelectric ordering in the ice crystallites resulting in a net dipole moment of the ice through molecular reorientation. This breaks inversion symmetry by proton flipping also inside crystallites (Fig. 9 right) and thus the bulk of the crystalline D2O layer contributes to the SFG process, leading to the tremendous SF signal enhancement.
Fig. 9: (left) SFG spectra (linear plot and inset logarithmic) of 8 BL crystalline D2O ice on Ru(001) (red trace). Upon irradiation with 4.66 eV photons, the SF signal is enhanced by a factor of ~103 (blue). If both UV and IR pulse temporally overlap, a further increase is observed (green). (right) Temporal evolution of the resonance amplitude centered at 2285 cm-1 and 2435 cm-1, respectively, which involves a structural transformation from ice Ih to ice XI (ferroelectric ice) via proton flipping (inset) as a proposed mechanism which leads to the observed SF intensity enhancement.
In a second project, the group has studied the defect-mediated chemistry on metal ox-ides, in particular the UV-photoinduced dissociation of N2O on thin MgO films on Ag(100) . Using postirradiation thermal desorption spectroscopy (TDS), the UV wavelength and photon dose-dependent formation of N2 in conjunction with recombina-tive desorption of atomic oxygen is found, while the coverage of the parent N2O molecules is depleted. If the atomic oxygen is not completely removed by high temperature desorption, the reactive sites for subsequent N2O photoreduction cycles are blocked. On the contrary, investigating the reactivity of these photogenerated atomic oxygen species on MgO showed that CO oxidation can be achieved using UV photoexcitation of the prepared CO+O/MgO/Ag(100) system. Currently, electron energy loss spectroscopy is used in the vibrational fingerprint region to identify the reaction products directly on the surface and not only indirectly with TDS, where photo- and thermally induced reaction products cannot be distinguished.
|| P. Giese et.al., J. Phys. Chem. C 115, 10012 (2011).|
2.3 Complex Dynamics
Studies of complex dynamics in chemical systems have been performed by the groups of Alexander Mikhailov and Markus Eiswirth, which build on the tradition of research on nonlinear dynamics and spatiotemporal pattern formation established in the department.
2.3.1 Complex Nonlinear Dynamics in Biophysical Systems
Rapid development of experimental methods in recent years has led to essential progress in molecular biophysics. Today, one cannot only observe single biological molecules, but also dynamically monitor their functional behavior or investigate the functional responses of protein machines to specific perturbations. Furthermore, non-invasive high-resolution optical methods provide the possibility to observe how spatial distributions of chemical species are dynamically evolving within living biological cells, and to investigate how such complex spatiotemporal processes are linked to the principal operation aspects of a biological cell. In essence, molecular cell biology is increasingly approaching a status of quantitative science, with the understanding being so profound that specific predictions become possible and the tools to control and steer biophysical processes can be developed. While molecular biophysics has started with the attempts to understand actual biological processes, the attention now gets shifted to synthetic and constructive biology and also to the design of artificial systems, such as protocells, which would reproduce certain properties of real biological objects and living biological cells.
The Complex Systems Group headed by Alexander Mikhailov is actively involved in these developments. In the last two years, firm contacts with leading international centers involved in the experimental and theoretical research on molecular biophysics have been established. Here, one should particularly stress the importance of collaborations with the recently founded RIKEN Center for Quantitative Biology in Kobe, Japan, and its director, Prof. T. Yanagida. In a joint project, the group aims to substantially deepen the understanding of how molecular motors, such as myosin, generate forces ultimately responsible for muscle contraction. Joint investigations, employing coarse-grained elastic-network dynamical descriptions of proteins, have allowed us to interpret the important experimental observations, indicating that myosin and, possibly, other protein motors are acting as molecular strain-sensors, so that their conformational responses to mechanical perturbations are central for their operation. As has been shown, mechano-chemical conformational dynamics in myosin and kinesin is intrinsically nonlinear and behavior of such motors cannot be understood with the usual normal-mode analysis .
Continuing previous research for another molecular motor, hepatitis C virus helicase, the group has finally produced numerical simulations which for the first time for any molecular motor reproduce in a structurally resolved dynamical way the entire operation cycles of this extensively studied protein machine, including its interactions with the double-strand DNA . Proceeding further, conformational dynamics in several other helicase proteins has been analyzed in a comparative study.
Fig. 10: Numerical simulations of adenylate kinase (ADK). ADK is an enzyme catalyzing the reaction ATP + AMP 2 ADP. Right: Sequence of conformational transitions inside the turnover cycle. The protein includes three distinct domains: CORE (green), LID (blue) and NMP (light blue). In the free enzyme (A), both domains LID and NMP are open. Binding of ATP induces closing of domain LID (B); subsequent binding of AMP leads to closing of domain NMP (C). In the fully closed state C, catalytic conversion takes place. After that, first LID gets open (D) and then also NMP, restoring the original free state A. Left: Probability distribution in the plane of distances RLC and RLN between the centers of mass of the CORE and LID and CORE and NMP domains, respectively. From Ref. .
In a collaboration with Prof. R. Kapral, the head of the Chemical Physics Theory Group in the University of Toronto, the group has been working on the combination of the elastic-network coarse-grained descriptions for protein with the coarse-grained hydrodynamic description of the solvent. In this direction, important progress could beachieved. Incorporating available experimental data, catalytic turnover cycles of the enzyme adenylate kinase (ADK) in the presence of solvent have been reproduced . Figure 10 shows the principal conformational states of this enzyme inside its turnover cycle and the computed probability density in the plane corresponding to distances between the principal protein domains. The current joint research, where scientists from the National Central University of Taiwan are additionally participating, is intended to enlarge coarse-grained descriptions, allowing us to consider operation of membrane-based protein machines in the presence of solvent.
At the next level of structural biological hierarchy, complex nonlinear dynamics of genetic regulation and protein signal transduction networks have been investigated. In an international project supported by the Volkswagen Foundation, the aim is to understand the origins of extreme robustness of real biological network-based systems to local structural perturbations and distributed noise. Here, the group collaborates with the RIKEN Center for Developmental Biology in Kobe, with the universities of Kyoto and Hokkaido, and with the Department of Mathematics in the Arizona State University. The first results include the model design  of synthetic oscillatory genetic networks able to maintain a required oscillation period with the accuracy of one percent despite the twenty-percent random variations of regulatory interaction strengths or knock-out of individual genes. Model signal-transduction networks with robust time-programmed functional responses have been designed and statistically investigated.
Going further to multi-cellular organisms, complex nonlinear dynamics in networks of diffusively coupled chemical microreactors has been theoretically considered. Together with Prof. H. Nakao from the Department of Physics of the Kyoto University (now in the Tokyo Institute of Technology), a detailed numerical study and statistical analysis of self-organized Turing patterns in network-based activator-inhibitor systems has been performed . This work has attracted broad attention, since it clearly demonstrates the existence of rich self-organization behavior in biochemical networks.
||Y. Togashi, T. Yanagida, A. S. Mikhailov, PLoS Comp. Biol. 6, e1000814 (2010).|
||H. Flechsig, A.S. Mikhailov, Proc. Natl. Acad. Sci. USA 107, 20875 (2010).|
||C. Echeverria, Y. Togashi, A. S. Mikhailov, R. Kapral, PCCP 13,10527 (2011).|
||Y. Kobayashi, T. Shibata, Y. Kuramoto, A. S. Mikhailov, Phys. Rev. E (2011, in press).|
||H. Nakao, A. S. Mikhailov, Nature Physics 6, 544 (2010).|
2.3.2 Electrochemical Dynamics
The group of Markus Eiswirth, studying electrochemical systems, has concentrated on possible gains in efficiency under oscillatory conditions in fuel cells as well as on spatiotemporal pattern formation on electrodes.
In general, dissipation in a process is given by the product of the flux and the driving force; for an electrochemical reaction these correspond to the current and overvoltage. The efficiency is then just the difference between the theoretically achievable energy gain (thermodynamic limit) and the dissipation. In a nonlinear system, during forced or autonomous oscillations, flux and driving force can be out of phase (or even anti-phase), which lowers the average dissipation compared to the stationary state. This idea was tested for electrochemical systems. The experiments were carried out in cooperation with Hamilton Varela (University of Sao Paulo at Sao Carlos, Brazil, and Ertl Center for Electrochemistry and Catalysis, Gwangju, Korea).
For any oscillator with hidden negative differential resistance (HNDR) it could be shown that the efficiency always increases at the onset of potentiostatic oscillations, since the negative differential resistance (where voltage and current are anticorrelated) appears only on a sufficiently fast timescale, which is exactly the scale on which oscillations set in. For galvanostatic conditions, the situation is more complicated; the result crucially depends on the shape of the nonlinear oscillations.
As an experimental example, the oxidation of hydrogen containing trace amounts of CO (which is realistic for technical fuel cells in which the hydrogen is obtained from methanol) was chosen, with pure oxygen as oxidant.
The dissipation for oscillatory conditions was calculated by integrating over several periods and subtracted from the equilibrium enthalpy of the total gas flux through the chamber to obtain the efficiency. A typical result is reproduced in Fig. 11, together with the data for pure hydrogen (for which the constant-current and constant-voltage data are identical and stationary). The presence of CO led to a significant decrease in efficiency due to partial poisoning of the active surface. However, for oscillatory (constant-current) conditions the efficiency significantly increased again, almost reaching the levels of pure hydrogen.
Fig. 11: Comparison of the efficiency of a fuel cell operated with pure hydrogen and hydrogen containing trace amounts of CO. The former could only be operated under stationary conditions, while the latter also exhibited oscillations when drawing a constant current leading to a significant increase in efficiency. The inset shows the highly nonlinear nature of the oscillations.
Research on spatiotemporal self-organization was continued using formic acid oxidation on a Pt ring. A number of new patterns were obtained such as localized bursts, standing domains and oscillating 4-domain patterns. Mechanistic studies of electrochemical processes were carried out in close cooperation with the Department of Inorganic Chemistry, starting with water splitting on platinum single crystals.
The Ertl Center for Electrochemistry and Catalysis (Gwangju, Korea), of which Prof. J. Lee and Dr. M. Eiswirth are vice directors, has become operative in 2010. At the moment the center has just one laboratory, but a new building is currently being constructed, which will provide substantial lab space and infrastructure.