Pictures SFB616

 Project A2:

Spectroscopy of Electronic Excitations in the Course
of Chemisorption and Chemical Surface Reactions

 Chemical reactions on surfaces are the backbone of all processes in heterogeneous catalysis. One reaction embraces a whole set of reaction steps as adsorption, eventually dissociation during adsorption, diffusion and finally the product forming step including desorption. The rate of the reaction as a whole is determined by a set of these discrete reaction steps and their corresponding activation energies. For example the energy gain of a chemisorption reaction can be partly converted into kinetic energy of the adsorbate. This kinetic energy can be used to overcome the activation barrier of surface diffusion leading thereby to an enhanced reaction probability with other adsorbates. But the nature of energy transfer between the single reaction steps is nearly unknown.

As a first reaction system we investigated the interaction of hydrogen atoms with noble metal surfaces. On noble metals only atomic hydrogen species adsorb and subsequently recombine to H2 molecules. The single reaction steps are adsorption, diffusion and recombination and desorption. Three different types of surface reactions can be distinguished (Fig. 1):

  • 1. Langmuir Hinshelwood (LH) reaction, where two adsorbed atoms
    react in the course of diffusion process to H2 and desorb.
  • 2. Eley Rideal (ER) reaction, where one adsorbed atom is abstracted
    from the surface by an incoming gas phase atom.
  • 3. Hot atom Eley Rideal (HA-ER) reaction, in which an impinging atom
    samples a significant surface area before reacting with an adsorbate.

Fig 1: Three types of recombinative desorption reactions on top of a metal-insulator-metal device.

These three reaction types can be investigated separately by the eligible choice of temperature and surface coverage when the reaction starts. The main question of project A2 is whether a part of the chemical or kinetic energy of the reactants is transferred into electronic excitations of the metal surface. This energy can be expected to be lower than the recombination energy of gaseous hydrogen (4.5 eV). Since the work function of noble metals is higher than this value, the electronic excitations cannot be observed as an external electron emission. But these electron hole pairs can be detected as tunnel currents in a thin film tunnel device with a tunnel barrier which is clearly lower than the work function of the metal.

Fig 2: Chemically induced tunnel current during two bunches of hydrogen atoms.

We use Ta/TaOx/Au systems with a tunnel barrier of 1.7 eV to detect the chemical induced excited carriers. These metal-insulator-metal (MIM) detectors are prepared by evaporating 30 nm thin tantalum films on flat glass slides followed by an electrochemical oxidation (resulting oxide thickness 3.5 nm) and by evaporating a 15 nm thin gold elektrode on top of the tantalum oxide. As a hydrogen atom source we use a radiation heated tungsten capillary for the dissociation of a hydrogen molecular flux. This kind of source emits thereby no ions and tungsten atoms. The source can be chopped. A typical signal of the sample is shown in Fig. 2 where the chemically induced tunnel current is shown while launching two bunches of hydrogen atoms (duration of each 20 s) on the gold surface of the device. A tunnel current of 214 nA can be monitored during an hydrogen flux of 1.4 · 1013 atoms/cm2s. This means that the conversion rate is 10-5 hydrogen atoms per electron.

Fig 3: Hydrogen induced tunnel current as function of hydrogen flux.

The tunnel current is proportional to the hydrogen flux in the bunches, this is plotted in Fig 3. Over a wide range from 0.34 · 1013 to 1.4 · 1013 atoms/cm2s the tunnel current increases linearly with the flux. When the hydrogen flux is switched off during a continued heating of the capillary, a small signal remains. This can be assigned to a photo induced current.

The chemically induced tunnel current may stem from electrons / holes which either propagate over or tunnel through the barrier. The height of this barrier can be changed during the chemical experiment by applying a bias voltage (see Fig. 4). A positive bias voltage decreases the barrier for electron tunnelling and increases the barrier for hole tunneling, while a negative tunnel voltage works vice versa. Tunnel junctions used in this project are generally two band tunnel devices. They allow a transport of hot electrons as well as hot holes. In fig.4 an example of a metal/insulator/metal junction is shown with a photoexcitation of the top electrode. Depending on the level of the oxides conduction and valence band one detects always a sum of two currents (hole and electron current) in the devices. While for an unbiased device the sum is negative and a net photoelectron current can be detected (left side of Fig.4), a biased voltage can show a net photohole current (left side of Fig.4). The influence of the bias voltage on a photoinduced tunnel current is shown for different photonenergies in Fig.5.

Figure 4: Energy level scheme of a metal/insulator/metal junction (y-axis: energy, x-axis: vertical cut of the system). Photoexcitation of the top electrode is shown as example.

Fig 5: Photocurrent (normalized with value at U=0 V) as function of the applied bias voltage. Calculated values for two different photon energies.

The calculated photoinduced tunnel current decreases amost linearly with the applied tunnel voltage for small photonenergies. For higher energies the decrease is exponential. This significant difference is used to characterize the energy distribution of chemically induced charge carriers in this project.

Future experiments in this project will touch electronic excitations during the oxidation of CO, the formation reaction of water guided by hydrogen atom and molecular beams on oxygen covered metal surfaces.





 This project is in collaboration with: A1, B3, C1



Internationaler Workshop 2008

Workshop 2008

A2 Schindler et al.
PDF (0.9 MB)

nternationaler Workshop 2008

Workshop 2008

A2 Stella et al.
 PDF (0.7 MB)

Remagen 2007

Remagen 2007
A2 Mildner et al.
 PDF (0.8 MB)