Search for spin-polarized photoemission from GaAs using light with orbital angular momentum

The Gay Group at the University of Nebraska wishes to improve state-of-the-art sources of polarized electrons which use circularly polarized light, and have recently studied the possibility of using CW (continuous wave) light with orbital angular momentum to produce photoemission from GaAs [1].

Post contributed by:
Nathan Clayburn, 
University of Nebraska-Lincoln, 
USA

A light beam with azimuthal phase dependence can carry orbital angular momentum (OAM) about its axis of propagation. This phase dependence corresponds to a helical path for the beam’s Poynting vector. Light with OAM is different from conventional circularly polarized light that possesses only one unit (±ћ) of spin angular momentum (SAM) per photon in that it can possess arbitrarily large values of OAM, ±mћ, where m is a positive integer.

GaAs has long been used to generate spin-polarized electron beams via photoemission using light with SAM [1]. These electron beams are created using atomically clean GaAs in an ultrahigh vacuum environment. Photoemission is obtained when chemicals (e.g., Cs and NF3) are applied to the surface to produce a negative electron affinity (NEA) condition. When circularly polarized light with near-band-gap energy illuminates the GaAs, spin-polarized electrons are emitted [3]. The degree of polarization, relative to an arbitrary axis of quantization, is defined to be
$$
P = { N_{\uparrow} - N_{\downarrow} \over N_{\uparrow} + N_{\downarrow}}
$$
with $N_{\uparrow}$ and $N_{\downarrow}$ being the number of electrons with spin up and spin down, respectively.

Figure 1: Energy level diagram for unstrained bulk GaAs. The values of $m_j$ are indicated below or above the respective states. The solid and dashed arrows indicate transitions caused by absorption of circularly polarized light with SAM of +1ћ and -1ћ respectively; the numbers associated with these transitions indicate their relative transition strengths.

The standard pumping scheme is shown schematically in Fig. 1, in which ~780 nm circularly-polarized light excites electrons from the conduction band to the valence band. The relative strengths for the two possible transitions (red arrows and blue arrows) dictate that the excited electrons will have electron polarization Pe = 50%. These electrons can then diffuse to the surface, where they “fall into the vacuum” by virtue of the negative electron affinity (NEA) surface of the GaAs. For unstrained GaAs, the theoretical maximum polarization is 50%, but in practice polarization is typically 25–35% due to various depolarizing scattering mechanisms.

Figure 2: Intensity patterns of various OAM
beams taken by the CCD camera (see text).
The integer number m is the topological
charge of the vortex. In the rightmost column
are the interference patterns obtained by
superimposing the positive and negative
OAM beams which reveals their “vortex” nature.

In our experiment, linearly polarized laser light was directed at diffraction gratings with screw dislocations to impart OAM to the laser beam [4]. The amount of the OAM and its’ sense of rotation was determined by the choice of grating. The light was then directed to a GaAs photocathode to produce polarized electron beams. For OAM to couple to a target, that system’s dimensions must have a characteristic length comparable to the spatial structure of the Laguerre-Gaussian transverse modes of the vortex light, shown in Fig. 2. We focused the laser spot tightly to reduce the length scale of its transverse modes in the hope of enhancing the angular momentum coupling to the lattice.  The electron beam (Fig. 3) was then delivered to a compact retarding-field micro-Mott polarimeter [5]. The polarimeter measures the counting-rate asymmetry in the elastic Mott scattering process, which is non-zero for spin states perpendicular to the electron scattering plane.

Figure  3. Schematic of the experimental apparatus used to measure the
polarization of electrons emitted by OAM light.

The polarization of electrons emitted from bulk GaAs due to absorption of 789-nm light with OAM = ±1ћ, ±2ћ, ±3ћ, and ±5ћ was measured and found to be consistent with zero with an upper limit of ∼3% (Fig. 4). This should be compared to Pe produced by using circularly polarized light of 32.32 ± 0.87(pol) ± 0.48(sys)%, where “pol” and “sys” refer to “polarization” (i.e., the uncertainty resulting solely from the Mott measurement) and “systematic” error, respectively. Our polarization measurements were consistent with zero, suggesting that light with OAM does not couple effectively to the internal motion of electrons in a semiconductor, at least when the focused transverse laser spot size is ∼200 μm or larger.

Figure  4. Electron polarization measurements for beams with various amounts of OAM. The composite error for the OAM polarization measurements was found by adding the systematic error associated with beam misalignment to the error of the electron polarization measurements using beams with OAM. This latter error includes a linear sum of the statistical uncertainty of the polarization asymmetry measurement and the systematic uncertainty of the polarimeter.

This technology development work was carried out in collaboration with colleagues at Jefferson Lab, and was funded by the National Science Foundation through Grant No. PHY-0821385 and under US Department of Energy Contracts No. DE-AC05-84ER40150 and No. DE-FG02-97ER41025.

References

[1] N. B. Clayburn, J. L. McCarter, J. M. Dreiling, M. Poelker, D. M. Ryan, and T. J. Gay Phys. Rev. B 87, 035204 (2013)
[2] D. T. Pierce and F. Meier, Phys. Rev. B 13, 5484 (1976).
[3] F. Meier, Polarized Electrons in Surface Physics, edited by R. Feder (World Scientific, Singapore, 1985), Chap. 10, pp. 423–466.
[4] A. M. Yao and M. J. Padgett, Adv. Opt. Photonics 3, 161 (2011).
[5] J. L. McCarter, M. L. Stutzman, K. W. Trantham, T. G. Anderson, A. M. Cook, and T. J. Gay, Nucl. Instrum. Methods Phys. Res., Sect. A 618, 30 (2010).