betaFeedback

WorksPhysics and Astronomy

Partial Quantum Coherence, Ultrashort Electron Pulse Statistics, and a Plasmon-Enhanced Nanotip Emitter Based on Metallized Optical Fibers

Abstract

The present dissertation covers two related research projects. The first topic was initiated with the ultimate goal of observing quantum degeneracy in ultrashort free electron pulses. This constitutes a thorough theoretical analysis of the problem involving partial quantum coherence and spin polarization of the source in light of a path-integral treatment of the phenomenon of matter-wave diffraction-in-time. Subsequently, results of a trailblazing experiment, to be superseded by a Hanbury Brown-Twiss type conclusive test of free fermion antibunching with electrons, is reported. In this experiment, the statistical distribution of the emitted electrons is studied taking advantage of a double-detector coincidence detection technique. The utilized electron emitters are ultrafast photoemission tungsten nanotip needle sources which are known to procure large spatial coherence lengths. The emission statistics is found to be sub-Poissonian. The second project involves introduction and full characterization of a novel laser-driven electron nanotip emitter based on metallized fiber optic tapers in which the emission mechanism is found to be assisted by surface plasmon resonance excitation as predicted prior to the design of such sources. It is shown that gold-coated fiber optic nanotips can emit electrons using both low-power continuous wave lasers as well as femtosecond pulsed lasers tuned to or near the surface plasmon resonance excitation wavelength of the system. The final chapter entails a proposal of a spin-polarized electron photoemitter based on the spin Hall effect for which such a fiber optic nanotip source is exploited.

by

A DISSERTATION

Presented to the Faculty of

The Graduate College at the University of Nebraska

In Partial Fulfillment of Requirements

For the Degree of Doctor of Philosophy

Major: Physics and Astronomy

Under the Supervision of Professor Herman Batelaan

Lincoln, Nebraska

The optically switchable spin-polarized tip source for the first project with the Pauli exclusion principle at its core is based on GaAs and is developed and characterized in-house by Mott polarimetry. Another novel spin-polarized emitter based on our proposed fiber optic nanotip sources therefore constitutes a confluence of the two research programs considered here and will wrap up the present dissertation. The reported findings in this work are of direct relevance to electron source and beam physics, ultrafast electron microscopy, and foundations of quantum mechanics in relation to the burgeoning field of free electron quantum optics.

dedicated to my mother and father,

with gratitude

Acknowledgements

I joined Prof. Herman Batelaan’s group in the summer of 2016. Looking back at the almost three and a half years which have passed since then, I feel enormously delighted and proud of being a member of this group and its achievements. In conjunction with that, I am deeply and graciously thankful of Prof. Batelaan for trusting me, investing in me, introducing me to multiple new things and laboratory techniques, giving me freedom to try and learn, and helping me shape my research career. I cannot overstate the professional support I received from Prof. Batelaan for which I will remain grateful the rest of my life.

Our experiments were performed in Prof. Kees Uiterwaal’s lab where we share the ultrafast laser systems. I would like to express my sincere gratitude to Prof. Uiterwaal for letting me work in his lab just like his own students. In addition, it has been a great pleasure for me to receive the opportunity to collaborate with Prof. Uiterwaal, and his graduate student Joshua Beck, on the project presented in chapter four of this dissertation. Moreover, I am indebted to Prof. Uiterwaal for reading my dissertation, and prior to that my comprehensive exam report, as a committee member; his comments on both the scientific aspects and the writing style of my dissertation have noticeably improved the entire work.

One of the projects presented in this dissertation has relevance to the research conducted on spin-polarized electron sources in Prof. T. J. Gay’s group. I received the opportunity to work in Prof. Gay’s lab in relation with this research for a time span of several months where I gained invaluable experience on this topic. I am extremely thankful of Prof. Gay for this, and for the helpful discussions through which I have learned a lot. In addition, as a committee member, I have to express my gratefulness towards Prof. Gay for reading my dissertation as well as my comprehensive exam report. In spite of all the possible deficiencies in my authorship of the present work, I can firmly assert that all the comments I received from Prof. Gay on how to present research work have improved my efficiency and prudence in delivering the results to the scientific community. I would also like to thank Prof. Gay’s students, Will Brunner for his collaboration, and William Newman and Karl Ahrendsen, for help during my time in Prof. Gay’s lab.

I am genuinely thankful of Prof. David Pitts for attending my presentations as the outside-representative committee member, and the instructive questions he challenged me with.

I shall express my extreme gratitude towards Dr. Ali Passian of the Oak Ridge National Lab with whom we had a fruitful collaboration on the project of chapter four of the present dissertation. It has always been a pleasure for me to communicate the results of experiments and data analysis we performed at UNL with Dr. Passian, and virtually always and every time, I learned a few new things from him. Dr. Passian’s positive attitude towards conducting professional research has become a great source of inspiration for me throughout the past few years.

I also thank Prof. Jeremy Armstrong from the University of Nebraska at Kearney for conversations on part of the project of chapter two.

I am thankful of Prof. Christian Binek for his support during my time at UNL before I joined Prof. Batelaan’s group, and for the valuable experience I gained and the lessons I took in his and Prof. Shireen Adenwalla’s lab.

I would like to take this chance and also thank my UNL friends for all their help and discussions on physics and beyond. Although I have to keep the list limited here, I genuinely need to express my indebtedness to Dr. Omid Zandi and Dr. Bahar Shahsavarani, both for all the discussions we had, and for their indispensable help during my transition time to the new environment where I had to learn how to pay the bills and get a driver’s license. I also thank Zilin Chen for discussions and sharing instruments, and Dr. Peter Beierle, and Dr. Maaneli Derakhshani, for fruitful discussions on foundations of quantum mechanics.

I am grateful of all the wonderful staff of the physics department, including the electronic and instrument shops. It was not possible to complete the experiments without their expertise. In particular, I would like to thank Bob Rhynalds, the instrument shop manager, and Mike Thompson, who nicely machined various crucial parts for my experiments, invariably starting off my rough wiggly hand-sketches.

Last but not least, without the research environment provided by the people and the University of Nebraska, none of the scientific achievements presented here would have been possible to deliver. In addition to this, I would like to express my gratefulness towards the Maude Hammond Fling Fellowship program at the university of Nebraska which I was honored to be a recipient of in the academic year 2018-19. I am thankful of the National Science Foundation which has funded the majority of the work presented in this dissertation through Prof. Batelaan’s research grant under the award number PHY-1912504. Part of the work was also funded by the Nebraska Research Initiative. I am grateful of the American Physical Society for granting me travel support to present our results in the APS DAMOP conferences twice.

I opened my acknowledgements by expressing my utmost gratitude towards Prof. Batelaan. I would like to point this out once more. I also add that as it goes without saying, in all the projects and their presented results, Prof. Batelaan’s ideas are implicit without my explicit mentioning of them. In spite of this, I emphasize that any possible deficiencies and shortcomings are only those of mine, and I take complete responsibility for the content of this dissertation.

Acronyms

AHE: Anomalous Hall Effect

ATE: Above-Threshold Emission

CC: Charge Current

CEM: Channel Electron Multiplier

CFD: Constant Fraction Discriminator

CM: Center of Mass

cps: counts per second

CW: Continuous-Wave

DIT: Diffraction-In-Time

ePPM: electron Point Projection Microscopy

FDTD: Finite Difference Time Domain

FE: Field Emission

FEQO: Free Electron Quantum Optics

FD: Fermi-Dirac (distribution)

FIB: Focused Ion Beam Milling

FN: Fowler-Nordheim (equation, emission, curve)

FONTES: Fiber Optic Nanotip Electron Source

GRIN: Graded Index

HBT: Hanbury Brown-Twiss (effect, interferometer)

HE: Hall Effect

HOM: Hong-Ou-Mandel (effect, interferometer)

ISHE: Inverse Spin Hall Effect

LED: Leading-Edge Discriminator

MCA: Multi-Channel Analyzer

MCP: Micro-Channel Plate (detector)

MCS: Multi-Channel Scalar

ML: Mode-Locked

MM: Multi-Mode

MPE: Multi-Photon Emission

MS: Mott Scattering

MWDIT: Matter-Wave Diffraction-In-Time

NIM: Nuclear Instrumentation Module

OAPM: Off-Axis Parabolic Mirror

O-FE: Optical Field Emission

OPA: Optical Parametric Amplifier

PEP: Pauli Exclusion Principle

PFE: Photo-Field Emission

PI: Path-Integral

PMT: Photo-Multiplier Tube

QE: Quantum Efficiency

QD: Quantum Degeneracy

QSHE: Quantum Spin Hall Effect

RLD: Richardson-Laue-Dushman (equation)

SC: Spin Current

SEM: Scanning Electron Microscopy(e)

S-FE: Schottky-assisted Field Emission

SHE: Spin Hall Effect

SNOM: Scanning Near-Field Optical Microscopy(e)

SO: Spin-Orbit (coupling, interaction)

SPES: Spin-Polarized Electron Source

SPM: Scanning Probe Microscopy(e)

SPR: Surface Plasmon Resonance

SQUID: Superconducting Quantum Interference Device

STM: Scanning Tunneling Microscopy(e)

TAC: Time-to-Amplitude Converter

TE: Thermionic Emission

T-FE: Thermally-assisted Field Emission

UED: Ultrafast Electron Diffraction

UEM: Ultrafast Electron Microscopy(e)

ULNES: Ultrafast Laser-driven Nanotip Electron Source

Outline of the main results

The main results to be reported and discussed in detail throughout the present dissertation, following the introductory CH. 1, are summarized below.

CH. 2

2.a The problem of quantum degeneracy in ultrashort free electron pulses, which is manifest through the fermionic Hanbury Brown-Twiss effect, is studied implementing a path-integral approach to matter-wave diffraction-in-time.

2.b In particular, realistic partially coherent pulsed electron beams are modeled using quantum mixtures of coherent and incoherent contributions to the total density matrix and for an arbitrary degree of spin polarization. The problem is also tackled in a more formal manner – though as a first step – based on quantum decoherence theory.

CH. 3

3.a An electron antibunching signal of 1 part in 4 in magnitude is reported on the detection plane using a femtosecond laser-driven tungsten nanotip photoemitter. This is significantly stronger than a previously reported electron antibunching signal of 1 part in 1000 using a high-brightness DC field-emission source from nearly two decades ago.

3.b Through an analytical calculation to characterize the sub-Poissonian nanotip source, accompanied by a computer simulation of the Poissonian counterpart of the same experiment, the electron-electron Coulomb repulsion in free space in ultrashort multi-electron pulses is identified as the antibunching mechanism. Other calculations show that we can approach the quantum degeneracy limit in upcoming experiments to study the Pauli force felt by free electron pairs.

CH. 4

4.a A new laser-driven nanotip electron source based on metallized tapered optical fibers is introduced, studied and characterized both optically and electronically. The emitter is back-illuminated and mechanically flexible.

4.b Fast continuous-wave electron emission is reported, and photo-field emission is identified as the underlying mechanism. The implemented laser powers are the lowest of their kind in the literature that have given rise to electron emission.

4.c Fast pulsed wavelength-dependent electron photoemission is reported using a femtosecond optical parametric amplifier as drive laser. Above-threshold emission facilitated by surface plasmon resonance excitation across a band in the visible spectrum is identified as the resonant emission mechanism. Multi-photon photoemission dominates the off-resonant domain.

CH. 5

5.a Spin-polarized sources of free electrons are directly relevant to the theoretical and experimental research work of CHs. 2&3. Given that ultrashort pulses of free electrons from coherent nanotip emitters are required to achieve enhanced antibunching signals investigated there, ultrashort spin-polarized sources of free electrons are to be developed. A proposal for one such nanotip source based on the spin Hall effect, which also exploits the fiber tip design of CH. 4, is put forward in the final chapter of this dissertation.

Prolegomenon to Part I (Chs. 2&3), Part II (Ch.4), and Their Confluence (Ch.5)

The present dissertation provides and discusses the results of the two major research projects which I worked on during my graduate studies in Prof. Herman Batelaan’s group at the University of Nebraska-Lincoln (UNL). The present chapter gives a general introduction to these two related lines of research, with both of which falling under the umbrella of free-electron physics: ultrashort free-electron second-order matter-wave interferometry, and its underlying technology, coherent laser-driven nanotip electron sources. Additionally, I will give an overview of the relevant basic concepts throughout this chapter.

Part I, to which the next two chapters are dedicated, covers the first subject. The theoretical analysis of quantum degeneracy (QD) in an ultrashort partially coherent spin-polarized free-electron matter wave and its characteristic antibunching signal, studied through incorporation of the fermionic Hanbury Brown-Twiss (HBT) effect [1-3], is given in detail in CHAPTER 2. While the analysis of the fully coherent pulsed beams is straightforward, that of the realistic partially coherent beams had not been reported before this research. In addition, the recent advances in making optically switchable ultrashort spin-polarized tip sources based on GaAs in Prof. T. J. Gay’s group here at UNL has been a further incentive to pursue this subject [4,5]. The reason is that as discussed in Ref. [3], observing the signature of the Pauli degeneracy force marks the prevalence of the quantum world where deviations from the apparent classical world are manifest. That is where electrons no longer seem to act like classical independent particles and therefore, their fermionic nature must be taken into account. A quantum mixture of independent ‘classical’ electrons and quantum mechanically correlated entangled electrons, described by an antisymmetric total wavefunction, leads to partial quantum coherence.

The study of free electron QD has its significance rooted in both fundamental and practical aspects. From the fundamental point of view, QD tells us when and how the underlying quantum world of the free electrons ceases to remain hidden under the guise of a stack of quantum mechanically uncorrelated ‘classical’ particles. When that becomes the case, the Pauli exclusion principle (PEP) must be accounted for, just like in the electronic orbitals of atoms where no more than a pair of electrons – and that with opposite spins – are allowed to occupy an identical orbital. We normally ought not to be worried about this effect for the beams of free electrons, that is, electrons propagating in free space, under the condition that they are born independently of each other with sufficiently long time-intervals between successive single-electron emission events such that multi-electron interactions are negligible. This is particularly true for continuously emitting sources of electrons like the widely used DC field emission tunneling sources as well as the thermionic emission guns. With the advent of ultrashort laser-driven nanotip electron sources (ULNESs) over a decade ago [6-9] and the large amount of research undertaken on this topic, it is now possible to generate ultrashort multi-electron pulses with a non-negligible chance to contain quantum-mechanically correlated electrons. QD then quantifies how likely it is for the electrons to fall within a single coherence volume, which is the jargon used to indicate whether the various relevant quantum mechanical effects (PEP, matter wave diffraction and interferences of various orders, etc.) can be observed or not. Beams of ultrashort free electron pulses thus seem to be providing a means for the experimental study of quantum foundations and photon-electron reciprocal control and interaction in the burgeoning field of free electron quantum optics (FEQO) [9], which can be seen as the electronic version of the deeply-researched field of photon quantum optics in free space, that has proved to be a trailblazer for quantum information science over the last two decades of the twentieth century. FEQO may someday culminate in reliably reproducible production of on-demand entangled free electron pairs. For the time being, we want to know how ‘quantum’ our ultrashort pulses of free electrons are.

The initial impetus of the research on ULNESs was primarily technological which provides a second reason why research on free electron QD has been of interest. The main advantage of nanotip sources lies in their significantly larger transversal (spatial) coherence lengths. In fact, the electrons in such point-like sources are generated within a very small volume with a radius on the order of 1-10 nm that encompasses only a few hundreds to a few millions of atoms. With their natural emission timing due to the well-defined repetition rate of the driving pulsed laser, these sources have been used in ultrafast electron diffraction (UED), as well as to make so-called molecular movies in time-resolved scanning electron microscopy (SEM) and transmission electron microscopy (TEM) schemes. Ultrafast electron microscopy (UEM) [10,11] is mainly performed in three modes of operation as discussed in Ref. [12]. In the stroboscopic mode, single- or few-electron pulses are used to form the time-resolved image over a relatively long experiment time. In the repeated mode, a larger number of electrons, on the order of 10410^{4} electrons per pulse, is used to construct the image at a faster pace compared with the stroboscopic mode, but at a cost of a reduced spatiotemporal resolution. This is due to the Coulombic interactions among the electrons of the bunch which is detrimental to the final resolution. In fact, such space charge repulsive effects not only broaden the pulse duration, but also contribute to the energy spread and beam divergence [13-15]. The loss of coherence in this case is more significant in the low-energy nonrelativistic regime where the typical interaction time is longer. For the more commonly used flat photo-guns, with a plethora of applications from electron microscopes to particle accelerators [16], the quantum coherence length and time is too short to make the Pauli degeneracy force a crucial factor to consider. For nanotip sources designed to give in a considerable spatial coherence, a conclusive characterization of the contributions of the various Coulomb and Pauli force effects has yet to occur. For the single-shot imaging and diffraction mode, which is the last of the main three modes of UEM, 10510^{5}-10810^{8} electrons per bunch is typical to achieve spatial and temporal resolutions on the order of 10 nm and 10 ns, respectively, although subpicosecond resolution is demonstrated to be within reach [12,15]. Here, in order to reduce the total exposure time as well as to gain a better contrast in the extracted images, a larger number of electrons are required to exist in each pulse, and based on the above discussion, understanding the electron repulsion effects both due to the PEP and the Coulomb interaction is demanded in order to fully characterize such multi-electron pulses.

These two incentives were the basis of the research that led to Part I of the present work. The detailed theoretical analysis of the former will follow in CHAPTER 2.

The first experimental results on this topic are provided and analyzed in CHAPTER § 10. Here, the experimental setup involves everything except for an electrostatic lens which is pivotal in exposing quantum effects by expanding the coherence volume through postselection of the coherently illuminated area. The significance of this experiment, however, is twofold. First, it is the natural precursor of the final free electron HBT experiments by ensuring that our coincidence detection electronics is indeed functional. It also turned out to be markedly helpful in regards with identification and subsequent suppression of the various noise sources and spurious effects before installing the electrostatic lens. All these proved highly advantageous in the design and construction process of a new high-vacuum chamber built in-house to hunt for the signatures of the Pauli degeneracy pressure in ultrashort free electron pulses. A discussion on the design of this new chamber and the experiment to follow will also constitute a portion of CHAPTER § 10.

The second equally important significance of the experiment whose results and analysis are elucidated in CHAPTER § 10 is that it provides a method to experimentally characterize the statistical randomness of the stream of electrons emitted from our source – a tungsten (W) nanotip needle. It is taken for granted in the community of electron physicists that the photoemission process by itself is statistically random, i.e. that which follows a Poissonian distribution for the consecutive emission time intervals [3]. In our experiment, we empirically inspected whether this is the case on the detection plane as well. Our method, based on utilization of a double-detector in a coincidence measurement scheme, is practical to deploy to characterize the particle statistics of virtually any free electron source, from photocathode emitters in various types of electron accelerators to ULNESs in UED and UEM applications and beyond.

Part II which is covered in CHAPTER 4, is on the development, physics, and characterization of a new electron nanotip source which works based on surface plasmon resonance (SPR) excitation of the system in the noble metal coating of tapered optical fibers using ultrafast drive lasers, and photo-field emission (PFE) using low-power lasers.

The significance of the nanotip sources was pointed out throughout the above discussion on the topic of Part I. It is therefore advantageous to characterize such laser-driven emitters. This is particularly true of the fiber-optic-based nanotip electron source (FONTES) presented in CHAPTER 4. The reason behind this emphasis is twofold. On the one hand, it is shown that such sources can be operated in the continuous-wave (CW) mode using low power CW lasers which can, interestingly, emit photoelectrons in the long wavelength regime, that is, with the photon energy of the driving laser being smaller than the work function of the metal. This is normally not expected for the nanotip sources which require illumination with femtosecond laser pulses that enjoy strong electric fields at each pulse peak. Such fields with appropriate polarizations are further intensified due to the geometrical shape of the metallic nanotip and can thus give rise to the nonlinear effect of multiphoton absorption for the electrons to overcome the potential barrier of the metal. A host of emission mechanisms including thermionic and thermally-assisted effects are also possible, at least in principle, which will be overviewed later in this chapter. To our knowledge, such a photoemission effect has not been reported for low-power (sub-mW) CW lasers.

The second reason why this new source is worthy of attention is that it can be operated in the pulsed mode as well. Although ULNESs in their modern form have now been around for nearly two decades, the FONTES proposed, studied, and characterized in this dissertation possesses a key advantage in contrast with the regular nanotip sources which is the alignment-free delivery of the emitted electrons anywhere inside the vacuum chamber. The only thing the user would need to do is to couple the laser light into the fiber conveniently from outside the vacuum chamber. The electron emitting nanotip at the metallized tapered end of the fiber can be positioned and oriented as needed and even raster scanned on top of a sample surface for potential time-resolved scanning tunneling microscopy (STM) applications. This is in stark contrast with how the common ULNESs are operated using alignment-sensitive optics to tightly focus the incoming laser beam on the nanometric emitting tip.

Despite being seemingly two separate research topics presented in a pair of distinct parts in three chapters, ULNESs are at the heart of this dissertation. In addition, as the PEP is a quantum effect that is intimately sensitive to the electron spin as will be discussed in detail in CHAPTER 2, spin-polarized electron sources (SPESs) are a key in segregating the effects of the Coulomb repulsion and that of the PEP for it is only the latter which responds to a ‘spin knob’. We use GaAs-based electron sources which can give rise to ultrashort electron pulses with a spin polarization of over 10% [4,5]. More details on this type of source will be given later on, but for now it should be clear why SPESs are important for FEQO aside from their significance in other fields of science and technology. CHAPTER 5 is a proposal for a novel SPES based on the spin Hall effect (SHE). The ultimate design is based on a FONTES coated with a material like W that can reportedly give rise to SHE. CHAPTER 5 is thus a confluence of Parts I&II.

In the remainder of this chapter, Section I.1 provides a more detailed introduction to Part I, followed by an overview of the fermionic HBT effect in Section I.2. Section II.1 serves as an introduction to the topic of Part II. Lastly, an illustrative overview of the electron photoemission physics and the role of the SPR in it when applicable provided in Section II.2 will close the present chapter.

Quantum coherence and the Hanbury Brown-Twiss effect

The observation of electron degeneracy pressure for a CW free electron beam with a high estimated degeneracy was reported in 2002 [3]. The free electron HBT effect was exploited to detect signatures of QD, and the observed characteristic feature, namely electron antibunching, has been claimed to be expressible by Coulomb repulsion rather than by electron QD in a subsequent theoretical analysis of the problem [17] which is also pointed out by the authors of a different article who performed a similar experiment using a TEM column [18]. Currently, the production of degenerate free electron beams is an open problem. The question on what limits the density of electron pulses, degeneracy pressure or Coulomb interaction, is both relevant for quantum mechanics and for applications such as making snap-shot movies of molecular interactions as pointed out in the opening to this chapter.

In the free-electron HBT experiment of Ref. [3], the nanotip electron source that produced a continuous emission of electrons, was claimed to have provided a world-record degeneracy of 1.6×10−41.6 \times10^{-4}, while femtosecond pulsed nanotip electron sources are projected to be capable of offering an electron degeneracy in excess of 10−110^{-1} per pulse. It will be discussed in detail throughout this and the next chapter that degeneracy in the context of unbound systems such as beams of free electrons is defined similar to that in bound systems – atoms and solid-state crystals, as representative examples. For instance, for fully occupied electronic states up to the Fermi level in a metal at a certain temperature, the system is fully degenerate below the Fermi level and non-degenerate elsewhere. Accordingly, for free electron beams, degeneracy is quantified as the number of electrons per coherence volume; the latter is defined by the longitudinal and transverse coherence lengths. We should also note that the concept of coherence in matter-wave interferometry is defined similarly to that in optical interferometry. It is therefore possible to characterize the coherence volume by, say, taking advantage of double-slit interference. Subsequently, from the average number of particles in a known beam segment, the degeneracy can be evaluated. For free fermions such as electrons, the maximum value of degeneracy is one in compliance with the PEP. The reason is that for electrons within a coherence volume, the only remaining quantum number that must differ for them to be able to occupy the same state, namely the same coherence volume, is the azimuthal spin quantum number and there are only two of them for the spin-1/2 electrons. For ultrashort electron pulses, the HBT antibunching signal, which, as will be shown, is a manifestation of QD, is therefore predicted in this research to be stronger by orders of magnitude compared with the CW experiment of Ref. [3]. Moreover, in a 1D approach, it has been shown that the Coulomb interaction and electron degeneracy both contribute to the free-electron HBT antibunching effect with equal proportions for pulsed nanotip electron sources [19].

A difficulty in the experiments is that the QD will not attain the maximum possible value of unity in practice. At such a value a fully coherent description of the quantum mechanical wavefunction would be adequate. However, at smaller values the electron sources and pulses are only partially coherent. To address the question how partial coherence influences the dynamics of two-electron wavepackets and their experimental antibunching signature, a theoretical framework is developed which will be discussed in detail in CHAPTER 2. We will conclude that the coincidence techniques can be used to attempt the observation of antibunching for particle-pair mutual detection, even when the electron pulse is partially coherent to the extent as expected for current pulsed nanotip electron sources. Our theoretical approach is intended to make estimates of the electron count rate and the coincidence signals as functions of time for design parameters of table-top experiments.

Historically, R. Hanbury Brown and R. Q. Twiss (HBT) demonstrated that intensity fluctuations of photons emitted from independent thermal light sources show positive correlations when two detectors are located close to each other; for detectors outside of the same coherence volume no correlations were observed [1,2]. The HBT stellar interferometer is shown in Fig. 1-1 schematically. The resultant photon bunching, which gives rise to an enhancement in the photon coincidence counts on two detectors within a single coherence volume, can both be justified by classical intensity fluctuation theory as well as quantum two-photon interference theory [20]. The former attributes the enhancement in photon counts to an increase in the normalized second order correlation function, which is defined as

g(2)=⟨I1I2⟩⟨I1⟩⟨I2⟩.(1)g^{(2)} = \frac{\langle I_1 I_2\rangle}{\langle I_1\rangle\langle I_2\rangle}. \tag*{(1)}

Here, I1I_1 and I2I_2 are time-dependent intensities detected by the two detectors, labeled 1 and 2, respectively, and the brackets indicate averaging over time. We will get back to the significance of this definition in section I.§ 9.4. The latter approach provides a full quantum optical treatment of the problem based on the bosonic nature of the photons [21].

Identical bosons tend to bunch due to the enhancement in their second order correlation function rooted in the two-particle interference with a positive sign, while identical fermions tend to antibunch because of the PEP which inhibits simultaneous phase-

Hanbury Brown-Twiss (HBT) stellar interferometer

Figure 1-1. Hanbury Brown-Twiss (HBT) stellar interferometer. The HBT interferometer was invented in 1956 by Robert Hanbury Brown and Richard Twiss. The main advantage of this interferometer over the then-common Michelson interferometer which used to be exploited to determine the angular size of distant stars is that, firstly, it is more compact. Secondly, by correlating intensities from independent detectors (rather than fields as in Michelson interferometer) it could bypass stringent stability requirements noting to the fact that the average intensity is not sensitive to atmospheric fluctuations. (a) Two possible two-photon trajectories, from (binary) stars aa and bb to two detectors labeled 11 and 22 located near each other on the Earth, are depicted by green solid arrows and red dashed arrows (more on this will follow in the discussion of Fig. 1-2). Intensity interference fringes are observed by varying the detectors separation dd. Here, the fringes constitute the number of mutual detection events at the detector pair in coincidence with each other labeled ncn_c (within an appropriately set window). This was justified by Roy Glauber in 1963 to be due to the bosonic nature of photons because of which the photons emitted from incoherent thermal sources tend to clump up (bunch) and arrive within a coherence volume in coincidence more often than a random stream of photons, that is, when the detector pair are separated by one coherence length. (b)

The effect is not restricted to double-stars. In fact, any pair of point-like sources on a single star can give rise to the HBT bunching signal. When integrated over the star surface, a similar effect will be observed. The angular size of the star can be determined this way when its distance from the Earth is known. This is explained later on in this chapter in the discussion related to Fig. 1-8.

space occupation of any coherence volume. The latter is indeed the case for a spatial wavefunction representing a two-particle interference with a negative sign. Strictly speaking, it is the symmetry of the spatial wavefunction of the identical particles under consideration that determines whether the corresponding statistical distribution should be bunched or antibunched. Clearly, for an electron pair prepared in a pure state with an antisymmetric total wavefunction, a spin-singlet state couples with a symmetric spatial wavefunction representing a two-particle interference with a positive sign. This indeed gives rise to an electron bunching effect. It is therefore important to note that particle bunching or antibunching per se are not inherent to the bosonic or fermionic nature of the streams of identical particles.

HBT-type correlations have been used to probe quantum mechanical particle statistics for various kinds of massive identical particles. Antibunching of thermal neutrons [22], antibunching in a Fermi gas of 40K{}^{40}\mathrm{K} atoms [23], and the HBT effect for 3He{}^{3}\mathrm{He} and 4He{}^{4}\mathrm{He}, with the former being a fermion (showing antibunching) and the latter a boson (demonstrating bunching) [24], were reported about a decade ago. The HBT effect with electrons in semiconductor devices typically consisting of electron sources, drains, and mesoscopic beam splitters was also observed [25-27]. In such experiments, high phase-space degeneracies, i.e. the number of electrons per phase-space-cell volume [19], are available and the Coulomb repulsion between electrons is screened by space charge effects. For free electron beams, as pointed out earlier, a conclusive demonstration of the HBT effect remains a challenge. A tour-de-force experiment was reported in 2002 [3]. A reduction of 1.26×10−31.26 \times10^{-3} in the coincidence count rate of free electrons in a continuous beam emitted from a high-brightness cold field-emission tip was demonstrated. However, at least one later analysis claimed that such small signals could have been dominated by Coulomb interactions between the free electrons in the beam, casting doubt on the original claim that a direct signature of the PEP had been observed [17-18].

Overview of the fermionic Hanbury Brown-Twiss effect

This section is meant to provide an introductory overview of the topic of Part I. Since the HBT effect employed in this research as a means of probing the quantum correlations in the ultrashort pulses of free electrons is an example of second-order matter-wave interferometry, where two independent detectors are used to probe the coincidences, an introduction to first- and second-order interferometry is given before discussing the free electron HBT effect as a specific example. Numerous sources, from books on quantum optics, quantum field theory, and quantum information theory, to research and review articles on both optical and matter-wave interferometry exist in the literature for an introduction to the topic under consideration here. Among them, Refs. [21,28-32] provide an inclusive list directly used in the following two brief introductory sections on the treatment of interferometric phenomena based on the first and second order correlation functions. Let us first juxtapose the first- and second-order interferometry schematically.

The two interferometry schemes are pictorially compared in Fig. 1-2. The well-known Young’s double-slit interference [33,34] which is a prime example of the more commonly taught and discussed first-order effect is depicted in panel (a). In optics, the first-order interferometry hinges on the interference of the electric field of light beams within a coherence length, while the second-order interferometry – perhaps counterintuitively at first sight – deals with the position-

Schematic comparison between the double-slit and HBT interferometry

Figure 1-2. Schematic comparison between the double-slit and HBT interferometry. (a) The phase difference between indistinguishable paths from the two slits which constitute the interfering alternatives gives rise to the first order quantum interference. (b) The general HBT setup involves two independent sources aa and bb, and two detectors 11 and 22 in the second-order interference. The green solid arrows and the red dashed arrows show the pair of indistinguishable two-particle trajectories as the pair of interfering alternatives. The general scheme of the figure is taken from Ref. [21]. (c,d) The two possible distinguishable combinations of two-particle trajectories which add up incoherently without contributing to the periodic interference pattern are depicted for comparison. These obviously do not form mutual alternatives. To observe this in a different way, as will be elucidated further later on in this chapter, for coherent electron pairs for which the two-particle state is required to be antisymmetric, the normalized total wavefunction in panel (b) becomes (φa1φb2−φa2φb1)/2(\varphi_{a1}\varphi_{b2}-\varphi_{a2}\varphi_{b1})/\sqrt{2} in terms of the single-electron wavefunctions each starting from one of the sources and striking one of the two detectors. This wavefunction is clearly antisymmetric under the exchange of any two source or detector labels. Considering the cases represented in panels (c,d), the hypothetical normalized wavefunction (φa1φa2−φb1φb2)/2(\varphi_{a1}\varphi_{a2}-\varphi_{b1}\varphi_{b2})/\sqrt{2} for the total state, would not be antisymmetric under the exchange of the detector labels. This shows that we cannot write down a total antisymmetric wavefunction for this pair of alternatives, just as

anticipated from the above-mentioned observation that they are indeed a pair of distinguishable two-electron paths which contribute incoherently to the final probability distribution – as opposed to contributing with an antisymmetric two-electron pure state like the indistinguishable alternatives of panel (b).

and time-dependent periodic pattern in the correlation of the intensity fluctuations of independent beams striking a pair of detectors. By correlation, we refer to the temporal average of the product of the intensities delayed by a fixed amount of time in space which can be measured experimentally in order to extract the corresponding oscillatory pattern. While in the double-slit experiment, the ‘which-way’ knowledge will wash-out the interference pattern formed by indistinguishable particle paths according to the standard interpretation of quantum mechanics [35,36], in a second-order interferometry experiment as in that of the HBT effect schematically shown in Fig. 1-2(b), it is the presence of a pair of indistinguishable two-particle trajectories that gives rise to the periodic pattern in the correlated intensities. The remaining two possibilities shown in panels (c) and (d) are of course distinguishable. To envision this, one may note that, say, in Fig. 1-2(c), the photon pair detected at the two detectors labeled 11 and 22 are both originated at a single point source bb; loosely speaking, the observer would only see one of the sources through her detectors which constitutes an unambiguously distinguishable ‘history’ as compared with panel (d) where the independent source aa is being observed. In Fig. 1-2(b), in contrast, there is no way to tell whether a photon registered at the detector 11 or 22 was originated at source point aa or bb so long as the detectors are coherently illuminated, i.e. positioned within one coherence volume of the incident particle stream.

First-order interferometry

Consider the positive and negative frequency components of the electric field operator denoted with E^(+)\hat{E}^{(+)} and E^(−)=(E^(+))∗\hat{E}^{(-)} = \left(\hat{E}^{(+)}\right)^*, respectively, with the former (latter) being proportional to the annihilation (creation) operator, a^\hat{a} (a^†\hat{a}^{\dagger}). The time averaged intensity of light at the space-time coordinate (r⃗,t)(\vec{r},t), say, on the detection plane is therefore dictated by ⟨E^(−)(r⃗,t)E^(+)(r⃗,t)⟩\left\langle\hat{E}^{(-)}(\vec{r},t)\hat{E}^{(+)}(\vec{r},t)\right\rangle. The first-order (amplitude) correlation function takes such terms at the same or at two different space-time coordinates, e.g. at the pair of pinholes in the Young’s interference of Fig. 1-2(a), to express their degree of mutual coherence:

G(1)(r⃗1,t1;r⃗2,t2)=⟨E^(−)(r⃗1,t1)E^(+)(r⃗2,t2)⟩.(2)G^{(1)}(\vec{r}_1,t_1;\vec{r}_2,t_2)=\left\langle\hat{E}^{(-)}(\vec{r}_1,t_1)\hat{E}^{(+)}(\vec{r}_2,t_2)\right\rangle. \tag*{(2)}

For such different space-time points, this gives the cross-correlation terms which form the periodic interference pattern. In the Young’s experiment, for instance, it can be shown that the average intensity at the coordinate (r⃗,t)(\vec{r},t) on the screen is given by

G(1)(r⃗,t;r⃗,t)=G(1)(r⃗1,t1;r⃗1,t1)+G(1)(r⃗2,t2;r⃗2,t2)+G(1)(r⃗1,t1;r⃗2,t2)+G(1)(r⃗2,t2;r⃗1,t1).(3)\begin{aligned} G^{(1)}(\vec{r},t;\vec{r},t)={}&G^{(1)}(\vec{r}_1,t_1;\vec{r}_1,t_1)+G^{(1)}(\vec{r}_2,t_2;\vec{r}_2,t_2)\\ &+G^{(1)}(\vec{r}_1,t_1;\vec{r}_2,t_2)+G^{(1)}(\vec{r}_2,t_2;\vec{r}_1,t_1). \tag*{(3)} \end{aligned}

We also make the observation that at fixed points in space this relation can be utilized to discuss the interference of radiation or matter waves in time.

The main mathematical properties of the correlation function are as follows:

References

No references were extracted for this paper.

Paper details

vlnc:W_wmcy9scp
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents