Highly Coherent Electron Beam from a Laser-Triggered Tungsten Needle Tip

We report on a quantitative measurement of the spatial coherence of electrons emitted from a sharp metal needle tip. We investigate the coherence in photoemission triggered by a near-ultraviolet laser with a photon energy of 3.1 eV and compare it to dc-field emission. A carbon nanotube is brought into close proximity to the emitter tip to act as an electrostatic biprism. From the resulting electron matter wave interference fringes, we deduce an upper limit of the effective source radius both in laser-triggered and dc-field emission mode, which quantifies the spatial coherence of the emitted electron beam. We obtain $(0.80\pm 0.05)\mathrm{nm}$ in laser-triggered and $(0.55\pm 0.02)\mathrm{nm}$ in dc-field emission mode, revealing that the outstanding coherence properties of electron beams from needle tip field emitters are largely maintained in laser-induced emission. In addition, the relative coherence width of 0.36 of the photoemitted electron beam is the largest observed so far. The preservation of electronic coherence during emission as well as ramifications for time-resolved electron imaging techniques are discussed.

Figure 1

Schematic of the experimental setup. (a) Illustration of one-photon photoemission (blue) and dc-field emission (green). Electrons from states around the Fermi level $E_F$ are excited by laser irradiation and emitted over the barrier, which is lowered due to the Schottky effect. At sufficiently high dc fields the barrier becomes narrow enough to permit direct tunneling through it, giving rise to dc-field emission. (b) Ultrahigh vacuum setup. A near-UV laser beam (photon energy of 3.1 eV) coupled into the chamber via a polarization maintaining fiber is focused onto the apex of a tungsten tip. Interference patterns are obtained on the microchannel plate detector after the electrons have passed a freestanding carbon nanotube (CNT) beam splitter placed in close vicinity to the tip. (c) Scanning electron microscope image of a CNT grown over a hole of a SiN membrane. (d) Sketch illustrating the formation of a virtual (or effective) source behind the tip’s apex as indicated by five exemplary electron trajectories (solid black lines). The diameter of the virtual source (vertical red full line) is substantially smaller than the geometrical source size (blue dashed line).

Figure 2

Electron interference fringes on the detector screen. (a) Laser-triggered electron emission at a bias voltage of $U_{\mathrm{tip}}=-41\mathrm{V}$. Without laser illumination no electrons are observed at this voltage. Scale bar is 1 mm on the detector screen. (b) dc-field emission ($U_{\mathrm{tip}}=-53\mathrm{V}$). Images are obtained by superimposing 200 individual images that have been corrected for slow linear drifts (see Supplemental Material [17]). A modulation of the fringe pattern along the CNT direction is also clearly discernible, arising from local distortions of the CNT and locally enhanced dc fields. Line profiles of the interference fringes are integrated perpendicular to the orientation of the CNT with laser-triggered (c) and dc-field emission source (d). The box in the right-hand inset indicates the $9.3{\mathrm{mm}}^2$ large integration area. The left-hand inset shows a larger detector image. At least 21 fringes in the laser-triggered mode and 35 in dc-field emission mode are visible as indicated with arrows. The slightly finer spacing in dc-field emission is due to the smaller electron de Broglie wavelength (see Supplemental Material [17]).