Introduction

During the past two decades, the development of micro- and nano-fabrication technologies has positively impacted multiple areas of science and engineering. In the photonics community, these technologies had numerous early adopters, which led to photonic devices that exhibit features at the nano-scale and operate at the most fundamental level of light–matter interaction [28, 39, 18, 29]. One of the leading platforms for these types of devices is based on gallium arsenide (GaAs) planar photonic crystals (PC) with embedded indium arsenide (InAs) quantum dots (QDs). The PC architecture is advantageous because it enables monolithic fabrication of photonic networks for efficient routing of light signals of the chip [26]. At the same time, PC devices have low loss and ultra-small optical mode volumes, which enable strong light–matter interactions. The InAs quantum dots are well suited for quantum photonic applications because they have excellent quantum efficiencies, large dipole moments, and a variety of quantum states that can be optically controlled [24, 3].

Currently, the development of these photonic technologies is geared mainly towards applications in quantum and classical information processing. The first proposals for quantum information processing using QDs in optical microresonators were developed more than a decade ago in the broader context of quantum information processing using quantum systems (such as atoms, ion, molecules) that can be optically controlled [23, 17]. Compared to other systems, the solid-state quantum photonic platform is attractive for quantum information applications because of its potential for large-scale integration [27].

We show in this review that the spin state of a single magnetic atom embedded in an individual semiconductor quantum dot can be optically probed. A high degree of spin polarization can be achieved for an individual Mn atom using quasi-resonant or fully resonant optical excitation of the quantum dot at zero magnetic field. Under quasi-resonant excitation, optically created spin-polarized carriers generate an energy splitting of the Mn spin and enable magnetic moment orientation controlled by the photon helicity and energy. Monitoring the time dependence of the intensity of the fluorescence during a resonant optical pumping process allows us to directly probe the dynamics of the initialization of the Mn spin. The dynamics and the magnetic field dependence of the optical-pumping mechanism shows that the spin lifetime of an isolated Mn atom at zero magnetic field is controlled by a magnetic anisotropy induced by the built-in strain in the quantum dots. The Mn spin state prepared by optical pumping is fully conserved for a few microseconds. These experiments open the way to full optical control of the spin state of an individual magnetic atom in a solid state environment.

Introduction

The ability to control spins in semiconductor nanostructures is an important issue for spintronics and quantum information processing. Single-spin detection and control is a key but very challenging step for any spin-based solid-state quantum computing device. In the past few years, efficient optical techniques have been developed to control the spin of individual carriers [34] or ensemble of nuclei [22] in semiconductor quantum dots (QDs).

Introduction

Colloidal semiconductor nanocrystals were the first model systems to evidence radiusdependent energy shifts of excitonic states caused by three-dimensional quantum confinement (for reviews see e.g. [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] and references therein). Besides the large tunability of optical emission wavelength, such nanocrystal quantum dots exhibit high quantum efficiency q and can nowadays replace organic dyes in various applications, e.g. act as biomarkers or active laser medium. In the broad field of nanoplasmonics, combinations of metallic nanostructures with semiconductor nanocrystals have potential applications in optoelectronics. The enhancement of spontaneous emission rate near metallic surfaces makes nanocrystals attractive candidates as probes of electromagnetic field distribution and nano-antenna effects. Charge separation at the metal–semiconductor interface can be applied in photocatalytic processes as was recently shown for production of hydrogen in multicomponent metal–semiconductor nanocrystal structures [11, 12]. The modification of optical and electronic properties of colloidal nanocrystals close to a metallic surface is a longstanding issue of research and will be reviewed in this chapter, in particular with a focus on the latest developments. We will start with an overview of fundamental properties of colloidal quantum dots along with a presentation of the most recent results in the field of functionalized colloidal nanostructures. We consider semiconductor and metallic nanostructures separately and describe their properties as individual building blocks for future complex nanosystems. In the following sections we deal with coupling schemes of quantum dots to metal surfaces, discuss practical applications in all-optical plasmonic devices and outline perspectives in quantum optics with surface plasmons.

In 1946, E. M. Purcell predicted that the radiative lifetime of an emitter is not an intrinsic property but can be modified by structuring the surrounding electromagnetic field [36]. By inserting a semiconductor quantum dot (QD) in an optical cavity, one can accelerate or inhibit its spontaneous emission. In the present article, we show that the QD spontaneous emission can be deterministically controlled to fabricate bright sources of quantum light.

QDs in cavities: basics, motivation, first demonstrations

Light-matter coupling

We note f the ground state of the QD and e its excited state. For a cavity mode close to resonance with the QD optical transition, we consider only the states with 0 or 1 photon in the cavity mode. The states ∣e, 0〉 and ∣ f, 1〉 are coupled through light–matter interaction, with a constant g, where hg = ∣〈e, 0∣ Ed∣ f, 1∣, with d the dipole of the optical transition e → f and E the electric field at the QD position.

Each of the states ∣e, 0 > and ∣ f, 1 > are also coupled to continua of states: continuum of the free-space optical mode, phonons of the semiconductor matrix, etc. [4]. Here, we consider only the coupling to the continuum of the free-space optical mode, related to the cavity losses, with a constant γc. When g << γc, the photon emitted by the recombination of an exciton efficiently escapes outside the cavity. The QD optical transition radiative recombination rate can be accelerated (Purcell effect) or inhibited.

Introduction

Quantum information technology promises to offer incredible advantages over current digital systems, allowing intractable problems in science and engineering to be tackled almost instantaneously through quantum computing, and unconditionally secure communication over long distances using quantum key distribution. Many schemes have been developed to implement quantum computing, including using linear optics [28]. The linear optical approach has proved popular due to the limited decoherence of photons with the environment, and accessibility of the components required for simple experiments. At the heart of an optical quantum computer, or extended range quantum key distribution using quantum relays or repeaters [14, 8, 24], lie entangled photons. The characteristics of the sources that create entangled photons, and their properties, are therefore central to realizing the full potential of such applications.

Quantum dots are one technology with which entangled light sources can be built [6]. Although first realised only relatively recently [49], they in principle offer key fundamental and practical advantages over other entangled photon sources. In the fundamental sense, quantum dots can be triggered, so that no more than one entangled photon pair is emitted at a time. This is in stark contrast to Poissonian entangled light sources [47, 27, 13], including the most widely used parametric down-conversion, where zero or multiple photon-pairs are usually emitted due to their probabilistic nature. Furthermore quantum dots have the potential to operate with high efficiency, with current experiments reporting up to 72% collection efficiency for the first and second photon [9, 12].

Introduction

The spin projections of single electrons and holes confined in quantum dots (QDs) provide a natural two-level system that can serve as the logical basis for both classical and quantum information processing devices. When two layers of self-assembled InGaAs QDs are grown sequentially, strain propagation causes QDs in the two layers to align along the growth direction. Coherent tunneling of either electrons or holes between the two QDs leads to a variety of Coulomb and spin interactions with possible applications in optoelectronic and logic devices [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Because coherent tunneling leads to the formation of delocalized molecular states, these vertically stacked pairs of QDs have come to be known as quantum dot molecules (QDMs). One of the surprising discoveries about QDMs was that the delocalized molecular states have their own unique and tunable properties [16]. In this chapter we review the formation of delocalized molecular states of holes in QDMs and consider how the structure and symmetry of the QDM influence spin properties. Results have been obtained for molecular states charged with one, two, and three holes [9, 13, 17]. We focus here on molecular states occupied by a single hole whose spin projections could serve as the basis for optoelectronic logic devices.

Hole spins were initially discounted for spin-based devices because the complex valence-band interactions were anticipated to degrade spin storage or decoherence times.

Introduction

Most of the current semiconductor devices rely on intentional, density- and spatially controlled doping with impurities. Dopants of donor and acceptor type enable both to change locally the electronic properties (conductivity, chemical potential, built-in electric field, etc.) and to tune these properties by metallic gates. Using such doping modulation has been shown to be very fruitful in the past two decades to fabricate and investigate semiconductor quantum dots (QDs) in the Coulomb blockade regime where the number of resident charges can be deterministically tuned one by one. In parallel, incorporating magnetic dopants in a semiconductor matrix has long been motivated by the possibility of inducing new properties and developing new functionalities. Observation of ferromagnetism in diluted magnetic semiconductors like Ga1–xMnxAs (with x in the range of a few percent) by the end of the 1990s has more specifically stimulated a lot of work [14]. Even though the Curie temperature of GaMnAs below ˜ 200K is likely to limit its potential use for applications, this compound still behaves like an ideal system to investigate the setup and control of Zener-type ferromagnetism in semiconductors, where the Mn atoms incorporated in the GaAs matrix provide both localized magnetic moments and free carriers. Combining the properties of quantum dots with those induced by magnetic doping is naturally an attractive track to explore, both to tailor new spin-based quantum properties and to investigate the fundamental interactions between carriers and magnetic impurities at the microscopic level. In this perspective, the limiting case of a single magnetic atom in a single quantum dot is obviously the elementary system of highest interest.

Semiconductor quantum dots (QDs) have been extensively researched in the past 20 years or so. Over this period, the field has been stimulated by various motivating factors from fabrication of low-threshold temperature-insensitive QD lasers to the use of single spins for quantum computing and single dots for medical markers. In the past decade, refinement of fabrication and experimental techniques enabled researchers in the field to routinely use single QDs to access and control single electrons and holes and their spins, and to generate non-classical light. The focus of this book is on control of optical and transport properties of single and few QDs. The remarkable progress in this fast-developing field in the past three to five years is reported.

The term “quantum dot”, widely used from late 1980s, usually refers to a semiconductor nano-structure. Typical sizes of a quantum dot range from a few nanometers in colloidal dots (also referred to as nano-crystals) to a few hundred nanometers in lithographically fabricated electrostatic structures, so that on average they contain from 103 to 106 atoms. The small physical size is the main common characteristic feature of quantum dots made from different materials and using various fabrication methods. It is usually combined with additional methods for electron energy engineering, for example, surrounding the dot with a higher band-gap semiconductor, applying gate-voltage creating a higher potential barrier around the dot, etc. This gives rise to the most important basic property of QDs: the motion of electrons and holes in QDs is suppressed in all three dimensions.

Introduction

The advanced growth of semiconductor quantum dots (QDs) with high optical quality is one key to realizing novel devices in various research disciplines related to modern semiconductor technologies. Despite recent progress in the top-down lithographic fabrication of single semiconductor QD-like emitters [43], bottom-up fabrication methods are commonly applied for the realization of high quality light emitting QDs [7]. The exploitation of high-density QD arrays as an active material in laser diodes [17] and vertical cavity surface emitting lasers (VCSEL) resulted in a new class of devices featuring lower lasing thresholds and improved device performance (such as temperature stability) in comparison to devices with higher dimensional gain [2, 3].

Spintronics and quantum information processing are intensively studied fields in order to provide complementary or entirely novel routes to a future information technology [12, 5]. Single self-assembled quantum dots (QDs) grown in low-density arrays are promising candidates for realizing functional building blocks in these research fields as they allow to confine single charge carriers or spin carriers while providing a solid-state platform capable of electrical injection or manipulation. Extensive research has resulted in the realization of various devices relying on few or single QDs, for instance quantum light emitters such as single photon sources [23, 36], sources of pairs of entangled photons [1, 45], and fewto single-QD lasers [38]. Moreover, many milestone experiments have been performed with In(Ga)As QDs grown on GaAs, for instance the observation of the Purcell effect [10] or reaching the strong QD exciton–photon coupling regime [26], demonstration of the indistinguishability of photons emitted by a QD [29], or complete control of a single spin by ultrafast optical pulses [25].

Introduction

This chapter gives an overview of our recent experimental research on graphene quantum dots. We will focus on two aspects, namely transport experiments revealing the existence of excited states and how they can be used to detect the electron–hole crossover and scanninggate microscopy to reveal local information on localized states.

One of the extraordinary properties of carbon atoms is the ability to be synthesized into solid structures of any dimensionality: three-dimensional graphite and diamond, one-dimensional carbon nanotubes, and zero-dimensional fullerenes. Graphene, a twodimensional carbon allotrope, was the last to be added to this list in 2004, when K. Novoselov et al. published their results [32]. Graphene has been stimulating the physicists' imagination since then: remarkable mechanical and electronic properties make it a promising candidate for future technological advances and breakthroughs in fundamental research [10]. In particular, the linear dispersion relation of graphene (for small energies) and an additional degree of freedom due to the valley degeneracy led to new physics not observed before in a condensed-matter environment. In fact, impressive results on the so-called unconventional quantum Hall effect [33, 55] and Klein tunneling [54] were obtained. The significance of the discovery of graphene and the first experiments were acknowledged and the discoverers of graphene were rewarded the Nobel Prize in physics 2010.

It was not obvious from the very beginning of graphene research whether it is easily possible to build more complex nanostructures like quantum point contacts or quantum dots (QDs).

Introduction

This chapter is devoted to the description of the interaction of polarized light with carrier spins and nuclear spins in semiconductor quantum dots. A historical starting point of these original experiments is the close analogy between quantum dot physics and atomic physics. In 1952, Brossel Kastler and Winter investigated mercury atoms in a weak magnetic field which splits the electron Zeeman levels. By irradiation of the atoms with circularly polarized light the authors could selectively populate one of the electron Zeeman levels [9]. This process has since been referred to as optical pumping. Soon afterwards the first optical pumping of carrier spins in a semiconductor was reported [31]. The initial pumping of spin-orientated conduction electrons in silicon induced by polarized light led to polarization of the nuclear spins of the atoms of the silicon lattice via the hyperfine interaction. This dependence of the nuclear magnetization on the polarization of the absorbed light is at the heart of the experiments described in this chapter. A review of the nuclear spin effects in bulk semiconductors can be found in [37]. The hyperfine interaction between carrier and nuclear spins gives even more spectacular results in quantum dots as shown in pioneering work on optically detected nuclear magnetic resonance ODNMR [23] and orientation of one spin species will have a strong influence on the other [25, 7]. Below we detail a selection of the most remarkable consequences of nuclear spin physics on the optical properties of quantum dots.