The Artemis ARPES end-station is designed for time- and angle-resolved photoemission spectroscopy. The end-station consists of a main mu-metal chamber for photoemission with a base pressure of 2x10-10
mbar. This chamber is equipped with a hemispherical electron analyser (SPECS Phoibos 100). The energy- and angle-resolved measurements are performed with a two-dimensional CCD detector, achieving an ultimate energy resolution of ~10 meV and angular resolution of <1°. Our current energy resolution is 130 meV, limited by the bandwidth of the XUV harmonic generated with a 30 fs pulse.
A liquid-helium-cooled, five-axis manipulator (azimuthal and polar angles) enables us to orient the sample crystallographic axis with the measurement plane and to cool the sample to 14 K (in <45 minutes) or e-beam heat it to 1000 K. The sample temperature is stabilised with a PID controller with a precision of a few percent of the desired temperature over a period of a few weeks. The surface crystallographic order is checked using a low-energy electron diffraction (LEED) analyser. A helium discharge lamp emitting at 21.2 eV enables off-line characterisation of the sample surface. This chamber is also equipped with a wobble stick for in situ sample cleaving.
The main chamber is connected both to the beamline (with a window valve) and to a sample preparation chamber. This second UHV chamber is dedicated to Ar ion sputtering, e-beam heating and thin film growth. Further flanges are available for users’ evaporators and sample storage capability is available. The sample is introduced to the preparation chamber through a fast load-lock chamber (sample turn-around time of 45 min) and then transferred to the experimental chamber with a transfer arm.
Time- and angle-resolved photoemission with XUV pulses
Time-resolved photoemission measurement on TaS2 with 17 eV photon energy.
Time and angle resolved photoemission spectroscopy (tr-ARPES) enables the electronic structure of a material to be monitored as it responds to excitation by a laser pulse. The target material is irradiated by a short laser pulse, which induces structural changes and excitations. It is then probed at a series of time delays by a short wavelength pulse which generates photoelectrons that are then collected and analysed.
Until recently, tr-ARPES measurements with lasers have typically used only near UV radiation (<7 eV) and pulses of 100 fs or longer. The low photon energy meant that only the region close to Gamma was accessible. The long pulse length meant that it was impossible to see the fastest changes to the material.
The Artemis beamline was one of the first in the world to overcome these limitations by using XUV pulses from high-order laser harmonics, with 20 eV photon energy and 30 fs time resolution. XUV pulses are created through the technique of high harmonic generation. A short pulse laser is focused into a gas-jet and interacts with the gas, producing even shorter pulses of coherent radiation in the 10-100 nm wavelength range. The higher photon energy enables electrons with a much wider range of energy and momentum space to be detected, meaning that each snapshot of electronic structure has a much wider field of view.
Our first time-resolved photoemission measurement with XUV pulses was performed on TaS2 at low temperature. The sample was pumped with 1.55 eV (800 nm) photon energy and probed with the 11th harmonic from the beamline. At zero time delay a clear collapse of the Mott gap was observed, characteristic of a photo-induced insulator-to-metal transition. The rigid shift oscillation observed at positive time delay is an amplitude mode of the charge density wave corresponding to the breathing of the Ta cluster. This work was published in Physical Review Letters (link opens in a new window).
We have now demonstrated both time- and angle- resolved photoemission measurements with XUV pulses on a number of materials, including graphene and topological insulators - see our publication list for details.
Ultrafast demagnetisation end-station
This UHV chamber is dedicated to time-resolved photoemission experiments aiming to detect very small variations of photoemission features. The technique implemented here is electron time-of-flight, where the photoelectron kinetic energy is deduced from the arrival time of the electron on the detector. This analyser is more suited to studying systems with low angular dispersion of the photoemission features such as core levels, or valence bands in some cases. The experimental chamber is equipped with a sputter gun, e-beam heating and LEED/Auger analyser for sample preparation.
Due to the time-of-flight technique, the noise level of this analyser is extremely good. First, the background electron count above the Fermi level is up to 3 orders of magnitude lower than the intensity on the valence band. This is particularly interesting for detection of small features in the empty states. Secondly, the high efficiency of electron detection and the low dark count (10-3 electrons/second) means that the noise is limited by the statistical number of electrons. For example, on the top of the valence band a pump-probe signal as small as 1% of the signal can clearly be detected making this technique appropriate for the study of valence band or core level photo-induced effects.
The energy resolution and angular acceptance depend strongly on the energy of the electrons in the drift tube, which is controlled by the retarding potential and the settings of the input electron optics. The geometric acceptance angle is ±1.5° and, using the XUV beamline, the best energy resolution expected is a few hundred meV.
In standard photoemission mode, an electron detector (MCP) is positioned at the end of the ToF drift tube to measure the spin-integrated photoemission. Connected to the monochromatised XUV beamline, this gives access to the valence band electronic structure and to the Linear Magnetic Dichroism measured at the M-edge of the ferromagnetic 3d-transition metals (Ni, Co, Fe). In-situ Magneto Optic Kerr Effect (MOKE) is available as a reference to characterise the macroscopic magnetisation. The chamber is equipped with pulsed magnetic coils to reverse the sample magnetisation between measurements.
Alternatively, the electron detector can be replaced (without breaking the vacuum) by a cylindrical electrode to transport the electrons to a spin detector (Mott polarimeter) to perform Time- Spin- and Energy- resolved photoemission simultaneously. This configuration is called the ToF-Spin mode. The analyser has been successfully used with ultrafast laser sources at 250 kHz repetition rate [refs 1-4 below] and 6.2 eV photon energy. We have just demonstrated operation at higher photon energy with the XUV beamline, which is technically challenging due to the 1 kHz repetition rate.
1 - C. Cacho et al., Spin-resolved two-photon photoemission on Fe77B16Si5 alloy (link opens in a new window), J Elec Spec Rel Phenom 169 62-66 (2009)
2 - C. Cacho et al., Absolute spin calibration of an electron spin polarimeter by spin-resolved photoemission from the Au(111) surface states (link opens in a new window), Rev. Sci. Instrum. 80, 043904 (2009)
3 - W. Wang et al., Fe t2g band dispersion and spin polarization in thin films of Fe3O4(0 0 1)/MgO(0 0 1): Half-metallicity of magnetite revisited (link opens in a new window), Phys. Rev. B 87, 085118 (2013)
4 - C. Cacho, et al., Momentum-Resolved Spin Dynamics of Bulk and Surface Excited States in the Topological Insulator Bi2Se3 (link opens in a new window),Phys.Rev. Lett. 114, 097401 (2015)