Femtosecond-Laser-Induced Spin-Polarized Electron Emission from a GaAs Tip
aa r X i v : . [ phy s i c s . op ti c s ] J a n Femtosecond-Laser-Induced Spin-Polarized Electron Emission from a GaAsTip
Evan Brunkow, Eric R. Jones, a) Herman Batelaan, and T. J. Gay Department of Physics and Astronomy, University of Nebraska-Lincoln, Lincoln, Nebraska 68588,USA (Dated: 10 January 2019)
It is shown that focusing circularly-polarized 800 nm light pulses of 100 fs duration on the tips of p -GaAscrystalline shards having no negative electron affinity (NEA) activation results in electron emission that isboth fast and spin-polarized. The 400 fs duration of the emission process was determined by pump/probemeasurements. The three samples we investigated produced electron polarizations of 13 . . . −
100 V and averagelaser power of 100 mW. The electron emission exhibited linear dichroism and was obtained under moderatevacuum conditions, similar to that of metallic tips. This source of spin-polarized electron pulses is “fast” inthe sense that the electron emission process is of comparable duration to the laser pulses that initiate it.
I. INTRODUCTION
Sub-picosecond, nanometer-scale, spin-polarized elec-tron sources are currently not available. Such a sourceis desirable for tests of quantum degeneracy and for ul-trafast electron microscopy.
The first reported obser-vation of free electron antibunching remains controver-sial, as the experimental apparatus could not distinguishbetween the effects of Coulomb pressure and degener-acy pressure.
As degeneracy pressure is polarization-dependent, while Coulomb pressure is not, a spin-polarized, sub-picosecond, nm-scale source could resolvethe controversy. The best combined spatial and tem-poral resolution in ultrafast electron microscopes is pro-vided by nanotip sources triggered by femtosecond laserillumination, as photocathodes with a planar geometryare limited in spatial resolution by the size of the laserfocus.
Direct measurements of the electron pulse dura-tion in ultrafast electron microscopy have shown that theelectron and the illuminating laser pulse durations are ofthe same order. Implementing a spin-polarized sourceinto an ultrafast electron microscope would allow for anovel approach to studying magnetic nanostructures atthe fs-scale. In this work, we present a fast, localized, spin-polarizedsource of electrons obtained from a sharp p -GaAs bulk[110] crystal shard illuminated with femtosecond laserlight. The size of the emission site is approximately 1 µ min scale, and the electron polarization achieved so far is13 %. The electron emission was studied using methodssimilar to those developed to characterize pulsed emis-sion from metallic nanotips. Such sources are currentlyin broad application to produce temporally short elec-tron pulses in beams with high brightness. Theyare referred to as “fast,” meaning that the temporal re-sponse of the emission process is comparable to that ofthe light pulse duration, and their spatial resolution has a) [email protected] been shown to be determined by the size of the emitterand not by the laser focus used.Standard CW polarized electron sources use a planarGaAs photocathode that must be layered with, e.g., Csand O to lower the vacuum potential below that of theconduction band. This creates a “negative electron affin-ity” (NEA) condition that allows electron emission by ab-sorption of a single photon from a CW laser (Fig. 1(a)).When circularly-polarized light with an energy near thebandgap ∆ of GaAs is used to excite electrons, there is animbalance in excitation probabilities of the two excited s / Zeeman substates (Fig. 1(c)), causing the emittedelectrons to be spin-polarized. Such sources are used ina variety of fields, including atomic and molecular, high-energy, and condensed matter physics. Alternative planar photocathodes with and with-out NEA have been developed to optimize the spin-polarization of the emitted electrons, to provide shortpulse operation, and to enhance source brightness. Back-illuminated NEA strained and unstrained thin photo-cathodes have produced 2 . There, the electron pulse duration is lim-ited by the slow emission process of diffusion throughthe material. A strained GaAs-GaAsP superlattice withNEA activation resulted in a 16 ps pulse duration ,and was used in a spin-polarized transmission electronmicroscope. The source was determined to have a de-generacy 2 orders of magnitude lower than the cathodetip used to first study free electron degeneracy, with asource size that was limited by the diffraction limit of thelaser focus. A planar GaAs photocathode with a Ag over-layer a few nm thick has functioned as a polarized elec-tron source without NEA activation by utilizing a multi-photon electron emission process. Electron yields wereincreased by employing local field enhancement throughplasmonic coupling on the surface of a p -doped GaAswafer, while the spin-polarization of emitted electronswas largely maintained. Pulsed ∼
100 fs laser light pro-duced a spin-polarization as high as 21 %, with a value of ∼
15 % for illumination at a central wavelength of 800 nm.Tips of magnetized iron and cobalt-coated tungsten E V p-GaAs Vacuum CBVB (cid:1) ϕ p-GaAs Vacuum CBVB
Cs, O (cid:0) E e - CsO e - σ + CB VB (cid:2) a) b)c) s p ħω FIG. 1. GaAs energy levels for (a) NEA bulk surfaces and(b) a non-NEA shard apex. The diagrams indicate bendingof both the valence band (VB) and conduction band (CB)at the surface due to heavy p -doping. (a) The vacuum en-ergy (dashed black line), is lowered (solid black line) due tothe deposition of alternating layers of Cs and O (top inset).Electron emission from the NEA surface proceeds by the ab-sorption of a single photon with energy that exceeds the bandgap ∆ of the bulk. (b) Multiphoton emission from an un-coated, non-NEA GaAs shard apex (see text). (c) Allowedtransitions at the GaAs Γ-point for absorption of right-handcircularly-polarized light by Zeeman ( m j ) sublevels. Selectionrules (∆ m j = +1) and the relative line strengths (indicatedin circles) yield a nascent conduction-band electron polariza-tion of (3 − / (3 + 1) = 50% for valence-conduction bandresonant transitions. have been used to produce spin-polarized electrons, al-though these sources have used only CW lasers todate. Such magnetized sources have a further lim-itation in that their spin polarization is not optically re-versible, unlike that of of GaAs photocathodes. An arrayof etched GaAs tips, illuminated with CW laser light forboth positive electron affinity (PEA) and NEA surfaceconditions resulted in a maximum polarization of 37 %,but the electrons were not pulsed. Implementation of atip geometry results in field enhancement at the tip apex,which increases the yield of emitted electrons. While amore robust activation surface of layered Cs and Te hasbeen demonstrated, a tip geometry, as well as a mul-tiphoton emission process, eliminates the need for NEAactivation that is sensitive to vacuum conditions. The work reported here focuses on obtaining fast, spin-polarized electrons from a sharp p -GaAs bulk [110] crys-tal shard, which naturally incorporates optical reversibil-ity. To do this, Ti:Sapph pulsed lasers with wavelengthscentered around 800 nm, the appropriate wavelength for single-photon excitation across the band gap, were usedto induce multiphoton emission without requiring thatthe samples have NEA. Fig. 1(b) illustrates this. Thevacuum potential (dashed black line) is modified at thesurface by the application of a negative DC bias voltage V and the local laser field (solid black line). A single pho-ton with energy just exceeding the bandgap ∆ can pro-mote an electron from the valence band to the conductionband. Absorption of a second photon can in principle re-sult in emission via tunneling through the vacuum poten-tial (blue arrow). Absorption of one or more additionalphotons provides sufficient energy for the electron to ex-ceed the additional ionization energy φ and escape intothe vacuum (red arrow). The 800 nm central wavelengthof our lasers accesses the relative excitation probabilitiesfor circularly-polarized light that make standard NEAGaAs sources produce polarized electrons (Fig. 1(c)). II. EXPERIMENT
We used two apparatuses, the first to measure elec-tron polarization and emission dichroism, and the secondto study the emission process duration and the emis-sion position dependence. Our first optical setup con-sisted of a Ti:Sapph oscillator (Griffin, KMLabs) withan output that passed through a collimating lens anda periscope placed prior to polarizing optics. A half-wave plate (HWP) followed by a linear polarizer wasused to vary the laser power without changing the di-rection of its linear polarization. The beam then passedthrough a quarter-wave plate to switch its polarizationfrom linear to left- or right-handed circular. A finalHWP was used to rotate the plane of polarization oflinearly-polarized beams. The laser then entered the po-larization/dichroism vacuum system through a window(Fig. 2). Just before entering the chamber, the width ofthe laser pulses was measured to be 75 fs with a SwampOptics Frequency-Resolved Optical Gate (FROG).The vacuum system, with a nominal base pressure of10 − Torr, comprised two sections. A sample chambercontained an off-axis front-surface Au parabolic mirrorto change the direction and focusing of the laser to a20 µ m-FWHM spot size. The GaAs shard was mountedon a 3-axis stage to position it in the laser focus. A chan-nel electron multiplier (CEM) near the sample monitoredthe electron emission current. We also measured the to-tal emission current from the electrically-isolated sample.Emitted electrons were directed to a compact, cylindricalMott polarimeter, comprising two concentric cylindri-cal electrodes and two CEMs placed symmetrically aboutthe entrance that defined the electron scattering plane.The central gold-plated electrode was biased at +20 kV,whereas the outer electrode and the mouths of the CEMswere biased at +500 V.To measure the electron polarization, P e , the countrates measured by the top and bottom CEMs ( C T and C B ) were monitored for electrons produced by light FIG. 2. The experimental setup for polarimetry and dichro-ism measurements. The pulsed laser beam (1) enters thechamber and hits the off-axis parabolic mirror (2) which fo-cuses the laser onto the sample (3). Note that the beam ispropagating out of the plane at (2), indicated by the red cir-cle. The sample is mounted on an XYZ translator (4) thatallows the sample tip to be positioned in the laser focus. ACEM (5) can be used to monitor electron emission. Trans-port optics (6) guide emitted electrons (7) toward the Mottpolarimeter (8) in the adjoining chamber with top (T) andbottom (B) CEM detectors. A 260 l / s turbomolecular pump(9) evacuates the chamber. pulses that were right-hand circularly-polarized, andthen compared with the rates when the light helicitywas flipped. The electron polarization, P e , is given as P e = S eff /A , where A = χ − χ + 1 and χ = s C T C ′ B C ′ T C B . (1)Here, S eff , the “effective Sherman function,” is the po-larimeter’s analyzing power, and the primes indicate theCEM rates for left-handed incident laser light. The ad-vantage of measuring P e this way is that it eliminatesfirst-order instrumental asymmetries. Measurements of the linear and circular emissiondichroism were made to better understand the emissionprocess. The dichroism, calculated using total emissionas measured by the CEM proximate to the sample, is D ≡ R − R R + R ; (2) R , is the rate of emission for orthogonal polarizations.Electron emission from the samples was optimized atthe edge of the crystal shard. Sharp tip-like shards weremade by shattering crystalline wafers and using an op-tical microscope to determine the “sharpest” pieces. When using these, total emission currents between 50 pAand 3 nA were obtained with an average laser power of ∼
100 mW and a DC sample bias of −
100 V.A second, similar apparatus was used to study emis-sion rates as a function of the shard apex morphology, to measure the dependence of emission rate on laser in-tensity, and to assess the temporal width of the emis-sion process. Pulses from a Ti:Sapph oscillator (SpectraPhysics Tsunami) were focused to a FWHM of 3 . µ m.The laser pulse intensity FWHM, τ laser , was measuredto be 100 fs. The laser power delivered to the shard apexwas controlled by a Brewster window variable attenuator.Pulsed electron emission was detected by a microchan-nel plate (MCP) placed close to the shard apex, or byan electrometer connected directly to the sample. Priorto entering the chamber, the primary beam was splitinto pump and probe components in a balanced Mach-Zehnder interferometer. The delay τ between pump andprobe pulses could be adjusted for values between ± τ > τ shows that the emission processdoes not exceed τ ; if τ ≈ τ laser , the emission processis “fast” as defined in the Introduction. Superadditivityfor τ ≫ τ laser implies the process is slow, e.g., due tothermally assisted processes. III. RESULTS
We first consider the electron pulse emission process.Electron emission from nanotips, if measured to be bothnonlinear and additive for τ > τ laser , has been shownto be fast.
Our electron emission current shows non-linearity as a function of intensity. It fits with a powerlaw of n = 5 . γ , characterizes the emis-sion. For γ ≫
1, field emission is dominated by multi-photon processes.
Given our focal spot sizes of 20 µ mand 3 . µ m, and an average power that never exceeded150 mW, our Keldysh parameter readily satisfied thiscondition in all our experiments and supports our simplemulti-photon model. The fifth order non-linearity indi-cates a five-photon process. (This result is in excess ofthe three-photon process illustrated in Fig. 1(b)). Gen-erally speaking, the order of the multiphoton process ina given sample can vary with the details of the emittingsurface, its local surface electric field, and the nature ofsurface states near the emission point. Pump-probe measurements as described above wereused to determine if the emission was additive.
Theadditivity ratio is defined as R ( τ ) ≡ R both ( τ ) − ( R pump ( τ ) + R probe ( τ )) R pump ( τ ) + R probe ( τ ) , (3)where R pump ( τ ) and R probe ( τ ) are the emission ratesfrom the pump and probe beams separately at each delay, (cid:3) m (ii)(i) a) n = 5.14(16) b) FIG. 3. Emission data from a GaAs shard “tip.” In (a), the additivity ratio R is plotted as a function of τ (blue circles). Thered line is the theoretical curve obtained with an electric field width of 160 fs and an I intensity dependence. The bifurcationof the R ( τ ) curve for τ <
400 fs is due to the flipping of the sign in Eq. 4 of E probe , and corresponds to the envelope functionfor the rapidly oscillating autocorrelation interference pattern in this region. The power dependence of emission is plotted inthe inset. (b) A scanning electron microscope (SEM) micrograph of the apex area with an expanded square section 20 µ m ona side. The laser focal spot size from the polarization measurements (dashed green circle, (i)) is shown to the scale of the topmicrograph and compared to the focal spot size (solid green circle, (ii)) used for the measurements shown in (a). Localizedemission from the shard’s sharpest features (inset) indicates that multiple sites may have been emitting in the polarization anddichroism measurements. and the rate R both ( τ ) was modeled as R both ( τ ) = ∞ Z −∞ [ E pump ( t ) ± E probe ( t + τ )] n d t. (4)The individual pump and probe field amplitudes weremodeled as Gaussians with E ( t ) = E exp[ − ( t/τ pulse ) ].The best fit to the data (red line in Fig. 3(a)) is obtainedfor τ pulse = 160 fs ( n = 5). The electron emission processis additive ( R ( τ ) = 0 for τ >
400 fs) and is thus shownto be faster than this value. Note that this is not a directmeasurement of the electron pulse duration. Neverthe-less, fast emission processes have so far indicated shortelectron pulses. We now turn our attention to electron polarization.Measurements of P e were taken with a 20 µ m-diameterfocal spot for two focal positions on the three sampleswe studied. In the first “tip” position, the focal spotwas centered on the shard apex. In the second “shank”position, the focus center was moved about 15 µ m awayfrom the tip towards the bulk. The results of all mea-surements of P e and emission dichroism, taken with the20 µ m focus, are given in Table I. In the “tip” position,with circularly-polarized laser illumination, P e was 13 %for samples 1 and 2, and 10 % for sample 3. Note thatthese results are comparable to those of Ref. 28. Varia-tions in the local structure or p -doping could be respon-sible for the differences in P e . As expected, when thelaser was linearly-polarized, the values of P e were con-sistent with zero. One exception, which we have yet to TABLE I. Polarization and dichroism results for circularly-and linearly-polarized light incident on either the apex (“tip”)or the bulk (“shank”) of three different shard samples.Target Light P e (%) D(%)Polarization understand, was observed with sample 2 in the shank po-sition. We note though that this value of P e is less thanhalf that of the polarization measured at the tip withcircularly-polarized light.Finally, we consider the sample morphology. The elec-tron emission rate was found to depend sensitively onthe position of the laser focus at the sample. Fig. 3(b)shows a plot of the emission rate measured in a 20 µ msquare area of a shard apex. The two laser focal spotsizes used in this work are shown relative to the size ofthe 20 µ m scale bar in the top micrograph. The bright-est emission feature was used to measure the emissiondependence on intensity (Fig. 3(a) inset) and to performthe pump/probe measurements.Non-zero linear emission dichroism (Eq. 2) was ob-served for the GaAs shards similar to a field emissiontip (FET). That is, emission is higher when the light’slinear polarization is parallel to the axis of the tip. In contrast, emission dichroism is absent for standardplanar GaAs sources.
Dichroism measurements weretaken at both focal positions as well. At the tip of theGaAs, the circular dichroism is small ( < in terms of nonlinearity, additivity, polariza-tion and local morphology although it is apparent fromFig. 3(b) that the overall shard morphology is complex.In summary, we have demonstrated a source that isable to produce fast pulses of polarized electrons froma micrometer-size area. This can, in principle, enablethe imaging of a small electron spot on a target to mea-sure spin-dependent effects with fs-scale resolution. Thereduced vacuum requirements of this source when com-pared with NEA GaAs sources make it easier and lesscostly to operate. Although the observed electron po-larization is modest, our results demonstrate that thissource follows the selection rules illustrated in Fig. 1(c).Polarization might be increased by having a sharper,more well-defined GaAs tip, or varying the laser wave-length. The parameter space is large and open to fu-ture study. Through the use of chemical etching and ionmilling, it is possible to shape the tip. An optical para-metric amplifier can be used to explore the wavelength-dependence of polarization. Investigation of the effectsthese parameters have on the total yield and polariza-tion of the emitted electrons is needed. ACKNOWLEDGMENTS
We thank M. Becker for useful conversations and S.Keramati for taking the electron micrographs shown inFig. 3(a). This work has been funded by NSF awardsPHY-1206067 and 1505794 (TJG), EPS-1430519 (HBand TJG), and PHY-1602755 (HB). V. A. Lobastov, R. Srinivasan, and A. H. Zewail,Proc. Natl. Acad. Sci. USA , 7069 (2005). P. Lougovski and H. Batelaan, Phys. Rev. A , 023417 (2011). E. Jones, M. Becker, J. Luiten, and H. Batelaan,Laser Photonic Rev. , 214 (2016). C. Ropers and M. Kasevich, personal communications (2018),femtosecond degenerate electron sources are currently being ex-plored at the University of G¨ottingen and at Stanford University. H. Kiesel, A. Renz, and F. Hasselbach, Nature , 392 (2002). T. Kodama, N. Osakabe, and A. Tonomura,Phys. Rev. A , 063616 (2011). G. Baym and K. Shen, “Hanbury Brown-Twiss interferom-etry with electrons: Coulomb vs. quantum statistics,” in
In Memory of Akira Tonomura (World Scientific, 2014) pp. 201–210. A. H. Zewail, Science , 187 (2010). R. Bormann, S. Strauch, S. Schfer, and C. Ropers,J. Appl. Phys. , 173105 (2015). M. T. Hassan, J. S. Baskin, B. Liao, and A. H. Zewail,Nature Photonics , 425 (2017). W. Wan, L. Yu, L. Zhu, X. Yang, Z. Wei, J. Z. Liu,J. Feng, K. Kunze, O. Schaff, R. Tromp, and W.-X. Tang,Ultramicroscopy , 89 (2017). B. Barwick, C. Corder, J. Strohaber, N. Chandler-Smith,C. Uiterwaal, and H. Batelaan, New J. Phys. , 142 (2007). P. Hommelhoff, Y. Sortais, A. Aghajani-Talesh, and M. A. Ka-sevich, Phys. Rev. Lett. , 077401 (2006). P. Hommelhoff, C. Kealhofer, and M. A. Kasevich,Phys. Rev. Lett. , 247402 (2006). C. Ropers, D. R. Solli, C. P. Schulz, C. Lienau, and T. Elsaesser,Phys. Rev. Lett. , 043907 (2007). D. T. Pierce and F. Meier, Phys. Rev. B , 5484 (1976). J. Kessler,
Polarized Electrons , 2nd ed. (Springer-Verlag, Berlin,1985). T. J. Gay, Adv. At. Mol. Phys. , 157 (2009). C. Prescott, W. Atwood, R. Cottrell, H. DeStaebler, E. L. Gar-win, A. Gonidec, R. Miller, L. Rochester, T. Sato, D. Sherden, et al. , Phys. Lett. B , 347 (1978). D. Androic et al. (Q weak Collaboration),Phys. Rev. Lett. , 141803 (2013). R. Feder,
Polarized Electrons in Surface Physics (World Scien-tific, Singapore, 1986). K. von Bergmann, M. Bode, A. Kubetzka, M. Heide, S. Bl¨ugel,and R. Wiesendanger, Phys. Rev. Lett. , 046801 (2004). F. Giebels, H. Gollisch, and R. Feder,Phys. Rev. B , 035124 (2013). K. Aulenbacher, J. Schuler, D. v. Harrach, E. Reichert,J. Rthgen, A. Subashev, V. Tioukine, and Y. Yashin,J. Appl. Phys. , 7536 (2002). M. Kuwahara, S. Kusunoki, X. G. Jin, T. Nakanishi,Y. Takeda, K. Saitoh, T. Ujihara, H. Asano, and N. Tanaka,Appl. Phys. Lett. , 033102 (2012). M. Kuwahara, S. Kusunoki, Y. Nambo, K. Saitoh, X. Jin,T. Ujihara, H. Asano, Y. Takeda, and N. Tanaka,Appl. Phys. Lett. , 193101 (2014). M. Kuwahara, Y. Nambo, K. Aoki, K. Sameshima, X. Jin,T. Ujihara, H. Asano, K. Saitoh, Y. Takeda, and N. Tanaka,Appl. Phys. Lett. , 013108 (2016). P. Klaer, F. Schertz, M. Lehr, G. Sch¨onhense, and H. J. Elmers,Phys. Rev. B , 214425 (2013). T. Irisawa, T. K. Yamada, and T. Mizoguchi,New J. Phys. , 113031 (2009). Y. Niu and M. Altman, Surf. Sci. , 1055 (2010). M. Kuwahara, T. Nakanishi, S. Okumi, M. Ya-mamoto, M. Miyamoto, N. Yamamoto, K. Yasui,T. Morino, R. Sakai, K. Tamagaki, and K. Yamaguchi,Jpn. J. Appl Phys. , 6245 (2006). J. K. Bae, L. Cultrera, P. DiGiacomo, and I. Bazarov,Appl. Phys. Lett. , 154101 (2018). M. Pirbhai, J. Knepper, E. T. Litaker, D. Tupa, and T. J. Gay,Phys. Rev. A , 060701 (2013). N. B. Clayburn, E. Brunkow, S. J. Burtwistle, G. H. Rutherford,and T. J. Gay, Rev. Sci. Instrum. , 053302 (2016). M. W. J. Prins, R. Jansen, and H. van Kempen,Phys. Rev. B , 8105 (1996). E. R. Jones, W. C. Huang, G. Basnet, B. N. Flanders, andH. Batelaan, Appl. Phys. Lett. , 263104 (2018). C. Kealhofer, S. M. Foreman, S. Gerlich, and M. A. Kasevich,Phys. Rev. B , 035405 (2012). F. V. Bunkin and M. V. Federov, Sov. Phys. JETP , 896. M. Kr¨uger, M. Schenk, M. F¨orster, and P. Hommelhoff,J. Phys. B , 074006 (2012). See, e.g., Section 3.1 of O. Schmidt, M. Bauer, C. Wiemann,R. Porath, M. Scharte, O. Andreyev, G. Sch¨onhense, andM. Aeschlimann, Appl. Phys. B74