6.1 Electrostatic lenses

6.1.1 Introduction

An electrostatic lens refers to a specially shaped potential with convex/concave features, similar to optical lenses, used to steer coherent electrons. The concept was first demonstrated experimentally in 1990 in [152, 153], in low-temperature, high-mobility semiconductors, which ensured that the coherent electrons had a sufficiently long mean free path to conduct experiments with structures made with the lithographic capabilities at that time. An experimental realization of the electrostatic lenses is illustrated in Figure 6.1.

The astounding decrease of the feature sizes in semiconductor devices, along with novel materials like graphene, has made (semi-)ballistic electron transport applicable at room temperature [154]. This has sparked new interest in applying electrostatic lenses in nanoelectronic devices, e.g. [155] suggests the use of lenses to focus electrons to the centre of nanowires, thereby avoiding rough interfaces and increasing mobility.


PIC

Figure 6.1: A GaAs/AlGaAs heterostructure, which forms a two-dimensional electron gas (2DEG) at the material interface. A potential is applied to a contact at the top of the structure, which is shaped to induce the desired potential barrier (electrostatic lens) in the plane of the 2DEG.


6.1.2 Law of refraction

Electrostatic lenses use analogous concepts from geometrical optics: Snell’s law describes the refraction of a light beam traversing an interface between two different media of propagation, e.g. air and glass. An equivalent law of refraction can be derived for electrostatic lenses using the principle of energy conservation.

A particle with a wavevector k has a kinetic energy

     ℏ2 |k|2
Ek = -2m-*-,                                                             (6.1)
where denotes the reduced Planck constant and m* the effective mass. As a particle traverses the interface between regions at different potentials (illustrated in Figure 6.2), its kinetic and potential energies change. The change in kinetic energy is attributed only to the change of the component of the wavevector normal to the interface (red); the component parallel to the interface (blue) is left unchanged. It follows that
|k |sin θ = |k |sin θ ,                                                         (6.2)
  1     1    2     2
where θ1 (θ2) is the angle of incidence (refraction) with respect to the normal of the interface. The magnitude of the wavevector is proportional to the square root of the kinetic energy:
               ∘E----
sinθ2 = |k1|=  ∘--k1.                                                         (6.3)
sinθ1   |k2|     Ek2
Therefore, the square root of the kinetic energy of a particle is analogous to the refractive index used in geometrical optics and this value can be dynamically modified by changing the value of the potential energy in the region of the lens. This concept is illustrated in the following section.

PIC

Figure 6.2: The wavevector of a particle is changed from k1 to k2 as the interface between regions at different potentials is traversed. The decomposition of each wavevector into its components normal (blue) and parallel (red) to an interface is shown. The normal component is modified according to the potential change, whereas the parallel component remains unchanged. The illustrated case assumes a positive potential step, where V 2 > V 1, such that the particle is refracted away from the normal to the interface.


6.1.3 Converging lens

Optical lenses which operate in a medium (air) with a lower refractive index use the familiar double-convex shape to focus light. Rays are refracted towards the normal of the interface as they enter the lens. The law of refraction for electrostatic lenses (6.2) reveals that if a positive potential step is used for the lens, the kinetic energy (refractive index) decreases and the trajectory of the electron is bent away from the normal to the interface. This dictates that a double-concave shape is needed to form a converging lens with a positive potential [153]. Conversely, a negative potential, i.e. a well, would require a double-concave shape.

The potential shape used to form the electrostatic lens is shown in Figure 6.3. The electrostatic lens has a peak potential energy of 40meV and the wavepacket is initialized with a kinetic energy of 180meV, moving rightwards. The electrostatic lens clearly focuses the wavepacket (density) after 150fs of evolution. A comparison to the evolution without a lens is shown in Figure 6.4, using a similar setup in the vertical orientation. A further consideration for designing the lens is that the size of the lens should be larger than the De Broglie wavelength of the electron (determined by its energy), otherwise an effective focusing will be distorted by diffraction effects.


PIC

PIC

PIC

Figure 6.3: The density of a Gaussian wave packet, over a sequence of time steps, is focused by a converging electrostatic lens with a double-concave shape (as indicated by the annotation in white). The wavepacket has a kinetic energy of 180meV and the lens has a potential energy of 40meV.



PICPICPIC (a) 0meV

PICPICPIC (b) 40meV

Figure 6.4: Wavepacket evolving freely (top sequence) and interacting with a double concave electrostatic lens (bottom sequence); the time steps (from left to right) correspond to 40fs, 100fs and 150fs.


The refractive index of the lens, and thereby its focal length, can be modified by varying the magnitude of the potential with which it is formed; a larger value of the potential energy for the lens (relative to its surrounding region) invokes a stronger refraction (’bending’). Figure 6.5 compares the effect of different values: It can be clearly seen that the barrier with the higher potential energy focuses the wavepacket more sharply (at the distance observed at the time instance shown). The applied potential can thereby control, when and at which distance, a wavepacket is focused (on a detector, for instance).


PIC (a) 10meV PIC (b) 20meV PIC (c) 40meV

Figure 6.5: Comparison of the density of a wavepacket evolved for 135fs in the presence of an electrostatic lens at various potential energy values to show different focussing.


6.1.4 Wavepacket control

The converging lens demonstrated in the preceding section, follows from classical analogues in geometrical optics. It is possible, however, to extend the concept of electrostatic lenses much further: engineered potential profiles can be used to dynamically control wavepackets, e.g. creating entangled states by splitting up wavepackets. It should be noted that the electron is not physically split; it is a single electron in an entangled state.

Figure 6.6 shows a rhomboid-like potential shape, along with the corresponding generation rate, which forms a lens that is able to realize such a function. The behaviour of the lens can be manipulated by changing the magnitude of the potential. This can be done dynamically, if the lens is realized with a structure similar to the one shown in Figure 6.1, using a time-dependent potential bias on the contact. Figure 6.7 illustrates the effect of the lens at different potential values. The density peaks indicate regions with a higher probability to find an electron. In Figure 6.7b (peak potential energy 70meV) the wavepacket almost fully traverses the lens and is split into two parts. The same lens shape, but with a potential energy of 120meV, splits the wavepacket into four parts (Figure. 6.7b): The front edges splits off a portion of the wavepacket by reflection, while the concave-shaped rear edges focus the transmitted parts again (applying the principle of Section 6.1.3). In the first case, with two peaks (the most-probable components of the state), the y-component of the wavevector remains positive, whereas for the second case, at a higher potential energy, the wavevector of the scattered state also has a negative y-component. This example clearly illustrates how specially shaped potentials can be used to influence the scattering pattern of an electron wavepacket. By varying the potential energy the electron can be guided in a certain direction with a controllable probability. This can be of use in the field of quantum computing to generate a (modifiable) entangled state and direct it to other computing elements.


PIC (a) PIC (b)

Figure 6.6: Two-dimensional potential (a) with rhomboid-like shape and concave-shaped rear edges, forming an electrostatic lens to scatter an electron wavepacket in various directions. The potential value of the lens is constant; it has no three-dimensional features. The corresponding particle generation rate γ is shown in (b).



PIC (a) PIC (b)

Figure 6.7: Wavepacket is split either (a) into two or (b) into four parts, after 90fs evolution, by a rhomboid-like potential profile with concave rear edges with a peak potential energy of 70meV and 120meV, respectively.