On analogue-gravity models as seen by in-universe observers
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
Scott Lockerbie Todd
BAppSc(Nano)/BAppSc(Phys)(Hons) RMIT
School of Science
College of Science, Engineering and Health
RMIT University
October 2020
Declaration
I certify that except where due acknowledgement has been made, the work is that of the author alone; the work has not been submitted previously, in whole or in part, to qualify for any other academic award; the content of the thesis is the result of work which has been carried out since the official commencement date of the approved research program; any editorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethics procedures and guidelines have been followed. I acknowledge the support I have received for my research through the provision of an Australian Government Research Training Program Scholarship.
Scott L. Todd October 16, 2020
i
Acknowledgements
Attempting to acknowledge every specific person who was of importance in my journey as a PhD candidate—important in the production of this thesis—seems to be at best a daunting task, and at worst a near-impossible task. If you, the reader, are somebody in particular to whom I owe a debt of gratitude, and if your name does not appear in the acknowledgements to come, then I sincerely apologize for the oversight; the omission is, I promise, unintentional. So where to begin? I guess I’ll try to do this somewhat chronologically. While it should go without saying, first and foremost: thank you, Mum and Dad. It’s difficult for me to properly articulate the multitude of ways in which I am thankful to you, but at the very least—and of direct relevance to this thesis—I think I can safely say that I likely would not be the inquisitive person that I am today (and I likely would not be writing an acknowledgement section for a doctoral thesis) had you not put so much effort into attempting to sate my curiosity at a young age. Much love and thanks to the both of you. I think it only fair to thank all of the education staff—from those earliest of teachers who taught me in primary school to those academics still teaching me in the present—whose collective hard work played no small role in my path through life leading to this point. Despite not knowing any of them personally, and despite basically all of my interactions with them having been remote, I feel like it would be unfair to not thank the staff at the RMIT library: so to the RMIT library staff, thank you! I definitely appreciate all of the effort that must have gone into acquiring all of the documents that I requested over the course of my PhD once my own attempts had failed to bear fruit. I am deeply thankful to the RMIT physics department collectively. Thank you to the academic staff and my fellow students for all of the guidance—both academic and personal—that has individually and collectively been imparted to me in my time as a student at RMIT. In particular of the RMIT physics department academics, thank you to Professor Andrew Greentree, Professor Jared Cole, and Professor Salvy Russo: the three of you have certainly put up with more unpleasantness from me than most of the academic staff at RMIT, in spite of which you have all remained absolutely phenomenal pillars of support both academically and personally. For that, I thank the three of you deeply. To Daniel Stavrevski and Giannis Thalassinos: you are two of the best friends that I have ever had the pleasure of making. The pair of you have ensured that my time at RMIT has been enjoyable since day one of our undergraduate (actually, day zero in Daniel’s case), and in complete honesty, I genuinely do not think that I could have gotten this far without the two of you. Your friendship and comradery means more to me than I’ll probably ever be able to express. I thank you both deeply for the everything. Furthermore on the student-front, I would like to specifically acknowledge Akib Karim, David Ing, Josef Worboys, Sam Wilkinson, and Tommy Bartolo. I could write paragraphs of thanks to each of you... but I won’t. Suffice it to say, your collective support, guidance, and friendship has been of immense importance to me through my time as a PhD student, and I am thankful towards the lot of you! Along with Daniel and Giannis, you guys have provided much entertainment, many fun times, and a whole lot of food for thought. Sam, you get bonus props for being an awesome DM during our D&D sessions (and I use the word “awesome” both in the original sense of the word and its more modern-day usage here)! Thank you to Dr Jamie Booth for all of the interesting discussions, guidance, and endless supply of entertaining anecdotes and stories! It should go without saying (though I shall of course say it anyhow) that I owe an immense debt of gratitude to the supervisory team that has seen me through my PhD candidature. Of my supervisors, I owe a special debt of gratitude to my primary supervisor Professor Nicolas C. Menicucci who has been a phenomenal mentor in more ways than I can adequately articulate: thank you very much for your endless patience, your guidance, and your feedback; I am—without doubt—a much better physicist for it. Thank you also to Doctor Valentina Baccetti, who joined as a secondary supervisor relatively late on during the course of my PhD candidature, but whose input as both a supervisor and a co-author has been invaluable. Salvy, you already got a thanks earlier on, so don’t get greedy! Penultimately, thank you to Past Scott, who in a cruel twist of fate can never read this acknowledgement. Past Scott was far from perfect—Present Scott isn’t either, and I very much doubt Future Scott will be for that matter—however he did put in a good amount of leg work at the start of the PhD that made putting this thesis together much easier than it otherwise might have been (and even then, it was a struggle at times). In spite of his desire to procrastinate, Past Scott made a decent effort to write explicit derivations down for Future (now Present) Scott’s reference/use, and I thank him for that. Finally, though definitely not least, thank you to my wonderful partner Thamale! Your endless love, support, and fantastic sense of humour has kept me sane and has made this whole thing possible. My love for you is nothing short of Jupiter-sized! rrr
iv Summary
The contents of this thesis include several investigations into analogue-gravity models from an internal or in-universe perspective. The philosophy driving these studies is that operational notions of observation and measurement have been of crucial importance in the construction of our theories of physics to date, and this was especially true in the construction of Einstein’s theories of special and general relativity: despite this, so far very little work appears to have been done on investigating analogue-gravity models from the perspective of such observers. Specifically, within this thesis it is demonstrated that it is possible to construct devices whose measurements behave in an equivalent manner to the reference frames within Einstein’s theory of special relativity. In particular, the collective measurements of these devices would lead observers who use them (in-universe observers) to view their analogue-gravity universe as containing a sonic analogue to special relativity (where the speed of sound 2s plays the role of the speed of light 2 from special relativity). Of course, the sonic analogue to special relativity (sonic relativity) is not fundamental: to this end, this thesis also provides a demonstration of how certain experiments can be performed by in-universe observers to reveal the fact that their universe has a preferred frame of reference (the frame defined by their analogue-gravity medium). The existence of a preferred frame of reference reveals to in-universe observers that there exists a bigger universe (the actual universe) outside of their analogue-gravity universe, and that sonic relativity is not a fundamental description of reality in this bigger universe. Towards the end of this thesis, it is speculated that it may be possible to map a certain type of experiment that violates sonic relativity to a type of experiment that does not violate sonic relativity. Should this be true, studying analogue-gravity models that are not fully sonically relativistic may still provide us with insights into real relativistic systems.
CONTENTS
List of Figures xi
I Background
1 Introduction1 1.1 Structure of this thesis ...... 2 1.2 The grand scheme of things...... 3 1.2.1 Gravity: fundamental or emergent? ...... 4 1.2.2 A small detour: Semiclassical gravity ...... 7 1.2.3 The case for quantizing gravity...... 8 1.2.3.1 A glance at the field of quantum gravity . . . . . 10 1.2.3.2 One final and curious quantum gravity result . . 13 1.2.4 The case for not quantizing gravity...... 14 1.2.5 Other possibilities: questioning our postulates? ...... 16 1.3 Analogue gravity...... 17 1.3.1 Origins...... 17 1.3.1.1 Gordon’s optical-metric model...... 17 1.3.1.2 Unruh’s acoustic black-hole model ...... 18 1.3.2 Developments...... 20 1.4 Making sense of it all: the role of analogue-gravity models in the hunt for a microscopic theory of gravity and the aim of this thesis 22
2 Theory 25 2.1 Relativity...... 25 2.1.1 Tensors and Einstein summation convention ...... 26 2.1.2 Special relativity ...... 33 2.1.3 General relativity...... 39 2.2 Quantum physics...... 42 2.2.1 Basics of quantum physics...... 42 2.2.2 The quantum harmonic oscillator ...... 45 2.2.3 Scattering theory...... 49 2.3 Field theory ...... 51 2.3.1 Quantum field theory ...... 53 2.4 Analogue gravity...... 57 2.4.1 Effective spacetime metrics ...... 57 2.4.1.1 Geometrical acoustic metrics...... 57 2.4.1.2 Physical acoustic metrics...... 59 2.5 An interesting toy model...... 65 2.5.1 The Lagrangian...... 67 2.5.1.1 The continuum limit...... 68 2.5.2 Equations of motion ...... 70 2.5.3 Bringing it all together...... 72 2.5.4 Including curvature (and thus some analogy to gravity) . 74
vii II Analogue-gravity universes and in-universe observers: Pub- lished/Submitted for publication
3 Sound clocks and sonic relativity 79 3.1 Introduction...... 79 3.2 Approach...... 83 3.3 Simple sound clocks ...... 84 3.4 Sound clock chains...... 88 3.4.1 Calibrating clock separation...... 88 3.4.2 Synchronisation of clocks...... 91 3.5 Relativistic effects observed by stationary sound clocks ...... 94 3.5.1 Time dilation as seen by stationary observers ...... 95 3.5.2 Length contraction as seen by stationary observers . . . . 97 3.6 Relativistic effects observed by moving sound clocks ...... 100 3.6.1 Time dilation as seen by moving observers ...... 100 3.6.2 Length contraction as seen by moving observers ...... 102 3.7 Sonic relativity...... 104 3.8 Discussion...... 107 3.9 Conclusion...... 109
4 Particle scattering in analogue-gravity models 111 4.1 Introduction...... 111 4.2 In-universe observers...... 114 4.3 Aim and Approach...... 116 4.3.1 Aim...... 116 4.3.2 Sketch of our approach...... 117 4.3.3 Schematic of our scattering experiment ...... 119 4.4 Phonon scattering kinematics...... 121 4.4.1 Phonon scattering from internal particles ...... 121 4.4.1.1 The laboratory frame kinematics of phonon scat- tering from internal particles ...... 121 4.4.1.2 The co-moving in-universe-observer frame kine- matics of phonon scattering from internal particles123 4.4.2 Phonon scattering from external particles ...... 125 4.4.2.1 The laboratory frame kinematics of phonon scat- tering from external particles ...... 125 4.4.2.2 Physical meaning of the two solutions ...... 127 4.4.2.3 The co-moving in-universe-observer frame kine- matics of phonon scattering from external particles128 4.5 Scattering cross sections ...... 130 4.5.1 General cross section definition ...... 131 4.5.2 Cross section in the in-universe observers’ reference frame 133 4.5.2.1 Reaction rate...... 133 4.5.2.2 Flux...... 134 4.5.2.3 Total cross section...... 136 4.5.2.4 Differential cross section ...... 136
viii 4.5.3 Phonons, quantized external particles, and the interaction Hamiltonian in the laboratory frame ...... 138 4.5.3.1 Interaction Hamiltonian ...... 139 4.5.4 Internal particle cross section ...... 139 4.5.5 External particle cross section ...... 141 4.5.5.1 Initial and final states ...... 141 4.5.5.2 Cross section derivation ...... 141 4.5.5.3 Scattering cross sections in the laboratory frame . 144 4.5.5.4 Scattering cross sections in the co-moving in- universe observer frame ...... 146 4.5.5.5 Using Lorentz-violating sonic Compton scattering to determine absolute motion ...... 146 4.6 Discussion...... 148 4.7 Conclusion...... 149
III Work-in-progress and additional unpublished work
5 From sound clocks to the Lorentz transformation 155 5.1 Defining four-vectors using sound clocks ...... 155 5.1.1 Clock-readings ...... 156 5.1.2 In-universe observer measurement of velocity ...... 156 5.1.3 Parametrising coordinates in one frame in terms of coor- dinates in another, or how I learned to stop worrying and love the Lorentz transformation ...... 158
6 Lorentz violating scattering 165 6.1 Notation and terminology...... 165 6.2 The Lorentz symmetry...... 165 6.3 Scattering from fields with different relativities ...... 166 6.3.1 In-universe observers and external particles ...... 166 6.3.2 External observers and internal particles ...... 172 6.3.3 Sonic relativity, photonic relativity, and the standard-model extension...... 174 6.3.3.1 Analogue gravity and the standard model exten- sion in the laboratory ...... 180
7 Towards sonic atoms 183 7.1 Sonically-relativistic quantum detectors ...... 183 7.2 A toy model for a sonic atom ...... 184 7.2.1 General requirements for sonic atoms ...... 188 7.2.2 An approximately one to one analogue of an actual atom . 190
IV Retrospective and outlook
8 Conclusion 197 Main references...... 200
ix Other references (etymology, interesting facts, etc..) ...... 220
V Appendix
A General transition probability derivation 223
x LISTOFFIGURES
2.1 Coordinate systems are a useful abstract mathematical tool, but unlike the things that they describe, coordinates are not inherently physically meaningful...... 28 2.2 A one-dimensional lattice of coupled pendula...... 66 2.3 The path of a single pendulim...... 67
3.1 The operation of a sound clock...... 85 3.2 The time taken for a sound clock to tick...... 86 3.3 Sound clock chains...... 87 3.4 Calibrating the seperation of sound clocks in a chain...... 90 3.5 Comparison of moving and stationary sound clocks: synchronisation of clocks and calibration of sound clock seperation...... 94 3.6 Measurement of the believed time taken for a moving sound clock to tick as measured from the reference frame of a stationary chain of sound clocks...... 96 3.7 Measurement of the believed seperation of pairs of sound clocks in a moving sound clock chain from the reference frame of a stationary chain of sound clocks...... 98 3.8 Measurement of the believed time taken for a stationary sound clock to tick as measured from the reference frame of a moving chain of sound clocks...... 101 3.9 Measurement of the believed seperation of pairs of sound clocks in a stationary sound clock chain from the reference frame of a moving chain of sound clocks...... 103
4.1 A scattering event as seen in (a) the laboratory frame and (b) the co-moving in-universe-observer frame...... 120 4.2 Phonon scattering from an external particle as viewed in the laboratory frame...... 132 4.3 Differential scattering cross sections for phonon scattering from external particles for fixed 0 and varying i of the external particle...... 147 4.4 Differential scattering cross sections for phonon scattering from external particles for fixed i of the external particle and varying 0...... 148
5.1 Two sound clock chains pass by one another in the laboratory, with the clocks at the origin of each chain being instantaneously synchronous.159
7.1 A diagrammatic representation of a Lorentz-obeying sonic atom. . . . . 187 7.2 A diagrammatic representation of a Lorentz-violating sonic atom . . . . 187
xi
Part I
Background
INTRODUCTION
The title of this thesis is, “On analogue-gravity models as seen by in-universe observers”; the term “analogue gravity” may or may-not be familiar to the reader, and so, for clarity, analogue-gravity models will be defined for the purposes of this thesis as such: 1 Analogue-gravity models are models of non-gravitational physical systems whose behaviour, in certain regimes, can be expressed in the mathematical language of general relativity.
An example might be illustrative here. Perhaps the most conceptually easy-to- understand and most well-known analogue-gravity model is that of the acoustic black-hole as proposed by Unruh in 1981 [1]. At the level of a verbal analogy, the idea is quite simple: a purely radially inflowing fluid that is allowed to accelerate as it flows will eventually flow faster than the speed of sound within the fluid itself; the region of the fluid at which the fluid-flow velocity is equal to that of the speed of sound in the fluid can be treated as an acoustic analogue to the event horizon of a black hole. Sound waves that are generated downstream of this acoustic-horizon will be swept downstream by the fluid flow quicker than they can travel upstream, never crossing to the region of the fluid on the other side of the horizon. This is—at least at the level of this verbal analogy—analogous to light’s inability to climb out of the gravitational well of a black hole. It turns out, however, that this analogy goes much further than merely a verbal one: as shall be discussed more heavily in Section 1.3, the mathematical formulation of this problem can be cast into a form that mirrors that of a massless field (such as light) propagating through a Schwarzschild spacetime in general relativity (GR), thus completing the promise made above when defining analogue-gravity models. While hydrodynamical analogue-gravity models are possibly the most well known—probably owed partly to their conceptual simplicity as can be seen from the acoustic black-hole example described above, and partly to the ease of their experimental realization (at least in comparison to some of the more exotic models)—plenty of other non-hydrodynamical analogue-gravity models do exist, and some subset of these others models will be discussed in Section 1.3 and beyond. However, before we can move on to these more interesting discussions, some housekeeping is in order: to this end, Section 1.1 provides the structure of this thesis.
1 2 Introduction |
1.1 Structure of this thesis
PartI –the current part of the thesis—contains two background chapters (detailed below) that are intended to be a relatively pedagogical introduction to analogue gravity and its related fields. If these chapters serve their intended purpose, then they should be of most use to students who are unfamiliar with analogue gravity and the fields from which it stems. While the chapters contained within this part of the thesis are far from comprehensive, they should ideally motivate analogue gravity as a research discipline, motivate the specific research topics of this thesis, and also serve as a relatively self-contained reference for future parts of this thesis.
Chapter1 —the chapter that you are currently reading—is a literature review that reviews, at a mostly non-technical level, the research en- deavour into unifying general relativity and quantum theory, analogue gravity, and its possible role in aiding this endeavour. Chapter2 provides the necessary mathematical machinery to make sense of everything that follows. The reader unfamiliar with the mathematical machinery of, in particular, general relativity and quantum field theory may benefit by reading Chapter2 before returning here to read the current one in full, as specific calculations in the literature will be displayed and discussed in Subsection 1.2.3 and Subsection 1.2.4. Alternatively, the reader can continue on bravely, consulting Chapter2 as necessary.
PartII contains two chapters, the content of which is work that is either published (Chapter3 ) or currently under review for publication (Chapter4 ).
Chapter3 is a minorly modified version of the first paper [2] produced during the undertaking of this PhD dissertation in which one method to construct operationally defined classical in-universe observers in analogue-gravity models is discussed. This paper was co-authored with Nicolas C. Menicucci. Chapter4 is a minorly modified version of the second paper produced during the undertaking of this PhD in which a toy model of particle scattering in analogue-gravity models is investigated.1 This paper was co-authored with Giacomo Pantaleoni, Valentina Baccetti, and Nicolas C. Menicucci, all of whom provided major contributions to Chapter 4.5.
1This paper is currently under review for publication with Physical Review D. A preprint version of this paper is accessible on arXiv [3]. The grand scheme of things 3 |
The appendix present in the paper was produced by Valentina Baccetti and is attached as AppendixA in this thesis.
Part III contains three chapters of unpublished work.
Chapter5 is an extension to the published work that makes up Chapter3 in which it is shown how to develop the full four-vector description of special relativity from the operational measurement procedure described in Chapter3 .
Chapter6 contains a rough sketch of where the work from Chapter4 may be taken in the future. In particular, it is demonstrated that it should be possible to leverage the Standard Model Extension (SME)—an extension to the standard model that allows for spontaneous Lorentz symmetry breaking—to further investigate such scenarios.
Chapter7 contains discussions and some preliminary back-of-the-envelope calculations on a particular type of sonically relativistic quantum detector.
PartIV contains as its sole chapter Chapter8 , which itself contains concluding remarks of this thesis and suggestions for where future research efforts may be best focussed.
PartV is the final part of this thesis and contains Chapter 4.5 as its sole entry, which as mentioned above is the appendix that was prepared by Valentina Baccetti for a paper that is currently under review with Physical Review D (the arXiv preprint can be found in reference [3]).
1.2 The grand scheme of things
To best understand the role of analogue-gravity models in the current landscape of physics, it is instructive to take a short wander through the history of the effort to unify general relativity with quantum theory. To that end, Subsection 1.2.1 will give a brief and high-level overview on the different conceptual avenues that are being considered to tackle this particular problem. The specifics of each avenue of research will be mostly delayed until Subsection 1.2.3 and Subsection 1.2.4, after which the reader should have a sufficient enough understanding of the research landscape to be able to understand the role of analogue-gravity models in the research endeavour, which shall be discussed in more detail in Section 1.3. 4 Introduction |
1.2.1 Gravity: fundamental or emergent?
Before moving on, it is important to explicitly explain what is meant by a fundamental interaction (alternatively, fundamental force, fundamental phenomena, or any other such terms formed by using the word fundamental as an adjective).
• Fundamental interactions are those physical interactions that cannot them- selves be explained in terms of other physical interactions. They are, in a sense, axiomatic (in that they do not follow from something prior). These interactions are intrinsic to certain physical objects, and so if the associated physical objects are present, then their corresponding interactions are also, by necessity, present.
As of the publication of this thesis in 2020, it is now widely believed that the physical laws that govern the universe are fundamentally quantum mechanical in nature. 2 The standard model of particle physics (SM)—whose mathematical framework is that of quantum field theory (QFT)—has had extraordinary success in explaining the observed properties of the material world3 to a high precision, specifically by providing a theoretical basis for the electromagnetic interaction, the strong interaction, and the weak interaction: three of the so-called fundamental interactions—or fundamental forces—of nature. Gravity is typically considered another fundamental interaction of nature bringing the total to four. While we have had an excellent classical theory of gravity—general relativity—for over 100 years, as of 2020, we have yet to establish an accepted theory of quantum gravity (QG). There have, of course, been many attempts to formulate such a theory , though as of present no candidate theory has been experimentally verified. It is entirely possible that future work may reveal one (or more) of these proposed theories to be a valid description of nature, but this remains to be seen.
2There are, of course, views to the contrary. Of historical note is the famous EPR-paradox [4], the suggested resolution of which was to accept that the wavefunction in quantum mechanics is an incomplete descriptions of reality and that there must exist additional elements of (physical) reality (or hidden variables), the observation of which would remove the indeterminate nature of quantum theories. The EPR-paradox, in-part, formed the basis for Bell’s famous paper [5] in which it was proven—under a reasonable set of assumptions about reality—that quantum mechanics cannot possibly be a local hidden-variable theory. It has since been suggested that our reality may in fact be consistent with a relaxing of Bell’s original assumptions, in which case Bell’s conclusions may be avoidable [6–10]. In recent years, ’t Hooft has investigated models of cellular automata (see [11] and references therein) as a means by which to describe quantum mechanics from a deterministic model: such models can be used to reproduce a 1 1 -dimensional quantum field ( + ) theory of non-interacting bosons [12]. 3By “material world” what is meant is matter and its associated non-gravitational interac- tions/behaviours. Gravity: fundamental or emergent? 5 |
Must there exist a theory of quantum gravity, though? Not necessarily. The strongly held view that such a theory must exist is based—as all non-axiomatic things are—on assumptions. There are certain fundamental assumptions that we make about the nature of the universe that, if invalid, would remove the supposed-need for a quantum theory of gravity. Two major assumptions that we typically make about our universe are the following:
1. Gravity is a fundamental phenomenon; 2. All fundamental phenomena are quantum mechanical in nature.
If one abandons the first assumption—that gravity is a fundamental phenomenon— then it must be the case that gravity is some kind of emergent phenomena (a definition of this term will be offered shortly). While certain emergent phenomena admit a meaningful quantum mechanical description (e.g., phonons and other condensed matter phenomena), others do not (e.g., thermodynamic properties like temperature). If gravity is not fundamental, and if it only exist as some kind of thermodynamic property of large ensembles of particles, then there may not exist any meaningful theory of quantum gravity. With that said, abandoning the first assumption does not necessarily lead to this conclusion: an emergent theory of gravity might still be meaningful before taking the thermodynamic limit (i.e., before all of the quantum mechanical degrees of freedom have been averaged out). It is therefore conceivable that we could abandon the first assumption and yet still be faced with the necessity of requiring a theory of quantum gravity. Abandoning the second assumption—that all fundamental phenomena are quantum mechanical in nature—is somewhat more controversial [13], though in principle this approach may actually be reconcilable with our current under- standing of the universe [14–18]; this shall be discussed in more detail below. Are there any reasons to actually believe that either (or both) of these as- sumptions might be invalid? In the case of the first assumption—that gravity is a fundamental phenomenon—the answer is both yes and no. Currently, there is no direct experimental evidence to support the notion that gravity is not a fundamental phenomenon, however, there exists both theoretical indications and a large body of indirect experimental evidence—by analogy—that gravity might be an emergent phenomenon. The specific arguments about the nature of gravity—fundamental or emergent—and the evidence to support these arguments will be relegated to the discussions in Subsection 1.2.3 and Subsection 1.2.4; before moving on to these discussions, however, we should discuss the notion of emergent phenomena in general in order to elucidate their possible role in explaining gravity. 6 Introduction |
The term emergent (and similarly related terms that are constructed by using “emergent” as an adjective) are often used fairly loosely, and so an attempt shall be made to make the idea somewhat more concrete for the purposes of this thesis. The notion of emergent phenomena/behaviour/physics/etc.. can be encapsulated by the following observation:
• The phenomenology of certain physical systems is not scale-invariant, with some phenomena often only appearing to exist for large values of certain physical parameters (usually the total number of particles in the system, or the physical size of the system).
That is to say, certain phenomena that did not appear to exist for small param- eter values appear to emerge at larger parameter values as some ensemble—or collective—property of the microscopic subcomponents (or degrees of freedom) of the system. Perhaps the two most familiar examples of emergent phenomena are that of thermodynamics—which emerges from statistical mechanics—and fluid dynamics—which emerges from molecular dynamics. A slightly more formal definition of emergent phenomena might be the following: emergent phenomena are those phenomena in a physical system that can be described in terms of macroscopic variables that are not inherent in the constituent components of the system. This then turns the question of, “what is meant by emergent phenomena?” into, “how does one define ‘macroscopic variables’?” One such answer to this question is the following: a macroscopic variable describes the average or net behaviour of some aspect of a physical system. A more specific definition might be the following: we choose to define a macroscopic variable to be any variable whose definition can be used to substantially reduce the number of degrees of freedom required to describe the behaviour of a physical system by—in essence—absorbing many individual microscopic degrees of freedom into a single variable whose value is insensitive to the specific values of any sufficiently small subset of those microscopic degrees of freedom from which it is defined. The prototypical example of a macroscopic variable is the temperature of some physical system in thermodynamic equilibrium: in this case, the definition of temperature absorbs a huge number of degrees of freedom—for example, the velocity of every individual molecule in an ideal gas—into a single scalar quantity. In keeping with the definition that was previously offered, the particular value of a system’s temperature is insensitive to a change in the kinetic energy of any sufficiently small subset of the constituent particles of that system (where in this case, by “any sufficiently small subset of the constituent particles”, we mean any number of particles that is small relative to the 1023 particles that one typically ∼ finds in a macroscopic sample of gas). A small detour: Semiclassical gravity 7 |
If it turns out that gravity—and perhaps our entire notion of spacetime—is some emergent phenomenon, then one should not necessarily expect to be able to naïvely quantize the objects that we use to describe gravity in general relativity (specifically, the metric tensor) because these objects might not exist in any meaningful way at the level of a microscopic theory. Whether or not quantization of the gravitational field would be meaningful within a theory of emergent gravity would depend on the specific nature of the emergent gravitational field itself. Before talking more about the methods that have been taken towards describ- ing the underlying nature of gravity (be it fundamental or emergent), let us first take a small detour to talk about semiclassical gravity.
1.2.2 A small detour: Semiclassical gravity
While the ultimate nature of gravity—fundamental or emergent—still eludes us, there does exist a semiclassical theory of gravity [14] whose validity we are relatively confident of in certain regimes. The semiclassical theory of gravity should exist as an effective mean-field theory in the correspondence—or classical— limit of any true quantum theory of gravity, and furthermore it should also produce valid predictions in certain limits of emergent gravity models. In the theory of semiclassical gravity, spacetime and its associated properties are treated classically, while the contents of the spacetime (e.g., matter fields) are treated quantum mechanically. Specifically, the gravitational field is treated as a classical field whose value is given as a function of the expectation values of quantum fields in spacetime. This is embodied mathematically in the semiclassical Einstein field equations [19]:
8 = ) #. (1.2.1) 24 h ˆ i
The spacetime structure (embedded in the Einstein tensor ) is given in its usual classical form on the left-hand-side, while on the right-hand-side we have the expectation value of the quantum energy-momentum tensor operator ()ˆ). The quantum energy-momentum tensor operator is sourced from some spatial distribution of matter described by the wavefunction # (note that # is a subscript on the right-hand-side, and so the wavefunction is not explicitly present in this form of the semiclassical Einstein field equations). While the physical regimes that would likely be required to test any full theory of quantum gravity or emergent gravity are currently far beyond our reach, the theory of semiclassical gravity should apply to physical scenarios that will be realizable in near-future experiments. For example, tangentially related to the 8 Introduction | semiclassical Einstein field equations is the Schrödinger–Newton equation [20–25]. The Schrödinger–Newton equation can be written as follows
2 %# 2 ¹ # C, A0 © ~ 2 2 3 ì ª 8~ C, A = < 3 A0 ( ) ® # C, A . (1.2.2) %C ( ì) −2< ∇ − ì A A0 ( ì) « ì − ì ¬ This is a modified version of the Schrödinger equation that contains, in addition to all of the usual terms, a potential term that takes into account the self-gravitation of some mass distribution that is represented by the wavefunction of the system. The gravitational field is described by a classical Newtonian potential, hence the name Schrödinger–Newton equation. Research on the implications of the Schrödinger–Newton equation constitute an example of research that is tangentially related to semiclassical gravity. This particular research avenue appears to be a promising one due to the fact that modern optical and optomechanical experiments should—in principle—suffice to determine whether or not the Schrödinger–Newton equation is valid within its regimes of supposed applicability. [26–28] The exact nature of the link between the Schrödinger–Newton equation and the theory of semiclassical gravity is ultimately reliant on whether or not gravity admits a quantum description. If the gravitational field is quantum mechanical in nature, then the Schrödinger–Newton equation is only an approximate description of the behaviour of large ensembles of gravitating particles; that is to say, in the limit of small number of particles, the Schrödinger–Newton equation fails to make valid predictions.4 If, however, the gravitational field only admits a classical description, and if it is also true that the semiclassical Einstein equations are indeed the correct way to describe the gravitational field of quantum mechanical systems, then the Schrödinger–Newton equation is valid even for single particles [25].
1.2.3 The case for quantizing gravity
Throughout this subsection, it will be assumed—unless otherwise explicitly stated—that gravity is a fundamental phenomena. Without experimental guid- ance, the current arguments in favour of quantizing gravity stem from demands of logical consistency; for example, by arguing that all fundamental phenomena must be quantum mechanical in nature, or conversely, by arguing that fundamental phenomena cannot possibly be classical in nature.
4This is demonstrated in reference [20] which appears to be the first instance of the Schrödinger– Newton equation in the literature (though the name was not coined until later, with reference [29] appearing to be its origin). The case for quantizing gravity 9 |
Perhaps the most obvious question that naturally leads one to realize that there should be a quantum description of gravity is the following: what happens when a quantum mechanical object is placed into a spatial superposition [30]? Taking our current understanding of gravity seriously, we observe that any object with mass/energy gravitates, which then leads very obviously to the conclusion that a physical object placed into spatial superposition should itself result in a superposition of gravitational fields (i.e., a superposition of spacetimes). This seemingly simple consideration appears to immediately necessitate the existence of a theory of quantum gravity.
The apparent necessity for a quantum theory of gravity can also be demon- strated by showing that physical theories involving both quantum mechanics and fundamentally classical gravitational fields lead either to internal inconsistencies or to predictions that cannot be reconciled with already understood physics. A particularly noteworthy example of this type of approach is the 1977 paper by Eppley and Hannah [13]: they argued that a fundamentally classical theory of gravity that is valid at arbitrarily small length scales results in violations of currently accepted and well tested laws of physics. In particular, Eppley and Hannah argued that if gravitational waves of sufficiently small wavelength are used to make measurements of quantum mechanical systems, then a violation of one of the following principles is a logical consequence: momentum conservation, the uncertainty principle, or locality. Eppley and Hannah’s proposal would be (by their own calculations) an engineering nightmare: the particular hypothetical experimental apparatus as envisioned in their paper would have an enormous mass, so much so that entire galaxies would be consumed in its construction. With that said, engineering nightmares are only problems of practicality: Eppley and Hannah’s calculations showed that the mass of their hypothetical experimental apparatus is less than the mass of the observable universe, and thus they con- cluded that their hypothetical experiment was at least possible in principle. The results of Eppley and Hannah would then seem to necessitate the abandonment of either momentum conservation, the uncertainty principle, or locality in any internally-consistent model of physics incorporating both quantum mechanics and a fundamentally classical gravitational field. Given how well these assump- tions have stood the test of time so far, the notion of abandoning any of them is difficult to accept.
It should be noted that Eppley and Hannah’s conclusions have subsequently been called into question [15–18]. Huggett and Callender [15] argued that Eppley and Hannah’s result is a consequence of choosing the standard (Copenhagen) interpretation of quantum mechanics and that alternative interpretations of 10 Introduction | quantum mechanics can—in an internally consistent manner—support a funda- mentally classical gravitational field. Even within the Copenhagen interpretation of quantum mechanics, the supposed problems that Eppley and Hannah predicted might actually be completely unobservable, and hence, unphysical. To this end, Mattingly [16] disputes the conclusions of Eppley and Hannah by demonstrating that their hypothetical experimental apparatus is, in fact, fundamentally impos- sible to construct in our universe even in principle. Through a more careful and thorough consideration of the necessary experimental considerations than was originally undertaken by Eppley and Hannah, Mattingly concludes that Eppley and Hannah’s hypothetical experimental apparatus would—by necessity of its design—sit within its own Schwarzschild radius. As a result, the apparatus would collapse into a black hole thus preventing its use as a measurement apparatus as intended. While Mattingly’s analysis rules out Eppley and Hannah’s particular thought experiment it does not necessarily rule out all possible thought experiments that could be designed to demonstrate the same thing. If it were possible to show that no experiment could ever be constructed (even in principle) to reveal the apparent physical inconsistencies that arise from coupling a classical gravitational field to quantum mechanical objects then, troubling though those inconsistencies may be, one could safely ignore them on the grounds that they have no measurable—and thus no actual—effect on reality. Mattingly briefly entertains such a scenario as being analogous to the cosmic protection hypothesis5 [31]6, dubbing it the semiclassical protection hypothesis. As of present, there does not—to the author’s knowledge—exist any no-go theorems that prevent some other hypothetical experiment from demonstrating what Eppley and Hannah’s specific experiment fails to. In the absence of any such no-go theorem, Eppley and Hannah’s result is certainly troubling. If the gravitational field is fundamental in nature then without abandoning or modifying quantum mechanics or other well tested principles of physics, quantization of the gravitational field may well prove necessary.
1.2.3.1 A glance at the field of quantum gravity
Many research programmes and theoretical frameworks have come into existence in the effort to find a theory of quantum gravity. Of those, the two largest and most noteworthy research programmes to date are that of Loop Quantum Gravity
5More commonly referred to as the “cosmic censorship hypothesis”. 6The original 1969 paper in which the cosmic censor was first proposed has been reprinted as [32]. The case for quantizing gravity 11 |
(LQG) [33, 34]—a theory of quantum gravity7—and String Theory8 (ST) [37]—a candidate theory-of-everything that itself contains theories of quantum gravity. Below is a short summary of the history of both loop quantum gravity and string theory. References to comprehensive review articles and textbooks on the subjects will be provided at the end of each summary. Loop Quantum Gravity is a direct continuation of the earliest efforts under- taken in the attempt to build a theory of quantum gravity. These first efforts took the obvious approach of trying to apply canonical quantization techniques to general relativity [38–40]. While these techniques had successfully allowed for the development of quantum mechanics and had paved the way for quantum electrodynamics (QED), attempts at canonical quantum gravity resulted in serious mathematical problems that stalled the research effort for some time. Eventually, work by Ashtekar [41, 42] allowed for the mathematical problems of canonical quantum gravity to be overcome and loop quantum gravity (LQG) emerged as a continuation of the canonical quantum gravity research programme. One particu- larly noteworthy aspect of loop quantum gravity is that it has been formulated in such a way that it is manifestly background independent, a feature that it shares in common with general relativity. Detailed discussions on the formalism and history of LQG can be found in Rovelli’s book [43] and the review articles [33, 34, 44], while references [45, 46] provide review articles for Loop Quantum Cosmology (LQC), which results from the application of LQG to cosmological scenarios. String theory has a rather more storied history than that of loop quantum gravity. Despite being a quantum theory, the path that leads to string theory in some sense begins with an attempt to form a classical unified theory of physics. Specifically, some of the major conceptual ideas underpinning string theory lie in Kaluza–Klein theory [47, 48], a theory that itself began as an attempt to “geometrize”9 electromagnetism. Originally, Kaluza demonstrated that by extending general relativity to include an additional fourth spatial dimension, both gravity and
7While loop quantum gravity itself is only a theory of quantum gravity, there do exist related theories sharing similar philosophical and mathematical bases that may prove to be candidate theories-of-everything. For example, Bilson-Thompson et al. [35] have shown that topological degrees of freedom within a loop-quantum-gravity like theory can be associated with the quantum numbers that describe particles. In particular, braided graphs within this theory can naturally be thought of as containing the first generation of standard model particles. 8Due to the vast number of theories that are contained under the umbrella term of “string theory”, it is somewhat difficult to point to single reviews on string theory as a whole. The resource letter in reference [36] contains a list of references that may be of use to the reader interested in specific aspects of string theory (e.g., bosonic string theory, superstring theory, M-theory, the holographic principle, AdS/CFT duality, etc..). 9Where by “geometrize” we mean “make geometric”, in the same spirit as to how “quantize” is understood to mean “make quantum”. 12 Introduction |
classical electromagnetism could be realized as a single geometric theory in a five-dimensional spacetime [49]10. A constraint referred to as the cylinder condition was imposed by Kaluza in order to suppresses the fourth spatial dimension such that the resulting theory more accurately agreed with our understanding of reality. Klein later demonstrated that the cylinder condition allowed for the electromagnetic field in Kaluza’s theory to be quantized if the fourth spatial dimensional was curled up—or “compactified”—into a cylinder with a radius of the order 10 30 cm [51, 52]. The notions of extra dimensions and dimensional ∼ − compactification from Kaluza–Klein theory eventually became central to modern string theory [47, 48] which originally emerged as its own theory during the period of time in which the standard model of particle physics was being constructed [53]. String theory was initially an attempt at describing hadrons [53], though ultimately quantum chromodynamics (QCD) succeeded in this endeavour by providing the mathematical formalism describing hadrons, their constituent components (quarks and gluons), and their interactions. Despite QCD’s success, string theory proved to have great theoretical utility in other ways: while not originally conceived of as a theory of quantum gravity, it was eventually realized that string theory naturally contains a quantized massless spin-2 field—the graviton—whose behaviour in the low energy limit is described by general relativity [54, 55]. In fact, not only does string theory allow for the existence of gravitons: it demands it. The existence of gravitons within string theory is, in fact, an unavoidable and necessary consequence on insisting that string theory is a unitary quantum theory, and for this reason—among many others whose detailing would make this introduction unreasonably long—string theory has received considerable attention by those in the quantum gravity community. See the 25th anniversary edition of Green, Schwarz, and Witten’s book for a foundation on string theory and for comments on how the field has progressed since its inception [56, 57]. Review articles on string theory (or particular aspects of it) can be found in references [58, 59], while reference [60] is a review article on String Cosmology, which results from the application of string theory to cosmological scenarios. Both of these research programmes have produced theoretical results that agree with those from the semiclassical gravity literature. In particular, both LQG and string theory admit derivations that correctly obtain results from black-hole thermodynamics, namely expressions for the Hawking temperature of black holes, and the black-hole entropy-area law (LQG references [61–63]; string theory references [64, 65]) . The fact that these theories make predictions that are in
10An English translation of the original German paper is available in reference [50] The case for quantizing gravity 13 | agreement with the predictions made within an existing theoretical framework whose validity we are reasonably certain of is a good reason to suspect that the research programme of quantum gravity is well founded. Loop quantum gravity and string theory are not—as has already been mentioned—the only quantum gravity research programmes. Other examples of research programmes and theoretical frameworks within the quantum gravity research endeavour include causal dynamical triangulation [66], twistor the- ory [67], and causal sets [68]. Some of the older and more established research programmes within the field of quantum gravity are touched upon by Smolin in reference [69] and Rovelli in reference [70]. This is not a complete list of all aspects of the quantum gravity literature by any means, and indeed such an endeavour is outside of the scope of this thesis. The literature does indeed appear to contain many compelling arguments and results that lend support to the idea that gravity should indeed be quantized. With that said, experiment is ultimately the arbiter of scientific validity, and to date, no experiments have been able to probe the physical regimes within which we expect the effects of quantum gravity to present themselves. While experiments that hope to probe semiclassical gravitational phenomena may provide us with some new experimental insight in the near future (as mentioned in Subsection 1.2.2), verification of the predictions made by semiclassical gravity will not necessarily allow us to uniquely identify the parent theory of which semiclassical gravity is an approximation. No theoretical framework for quantum gravity has yet been blessed with experimental validation, and so for the time being it is an open question as to which—if any—approach best describes the fundamental nature of gravity within our universe.
1.2.3.2 One final and curious quantum gravity result
As a final note on the quantum gravity literature, it is worth drawing attention to a curious result obtained by Dyson [71]. Assuming that there does indeed exist a theory of quantum gravity, Dyson asked the question “Is a graviton detectable?” in a paper of the same name [71]. While Dyson does not prove that gravitons are undetectable by any and all conceivable experiments, he does show that if one were to attempt to construct a device that follows the same design principles as LIGO (Laser Interferometer Gravitational-Wave Observatory), then in order for it to possess the sensitivity to detect single gravitons its pairs of mirrors would sit within their combined Schwarzschild radius. As a result, the device suffers the same fate as Eppley and Hannah’s device [16]: by necessity of its design, it collapse into a black-hole, preventing its intended use. Eppley and 14 Introduction |
Hannah’s hypothetical device was conceived of under the assumption that gravity is fundamentally classical, while Dyson’s hypothetical device was conceived of under the assumption that gravity is fundamentally quantum: in either case, the laws of physics—whether by coincidence or for some deeper reason—seem to conspire in just the right way as to prevent probing the fundamental nature of gravity. If the apparent inability to probe the nature of gravity turns out to be some fundamental limit that extends to any and all such experiments that might conceivably be used to do so, one may pause to wonder if perhaps Mattingly’s “semiclassical protection hypothesis” [16] is actually part of some more fundamental gravitational censorship hypothesis11 that makes it impossible to determine whether or not gravity is quantized. Such a prospect seems quite disturbing from the point of view of the theoretical physicist.
1.2.4 The case for not quantizing gravity
As has been discussed in Section 1.2, there exists the possibility that gravity is an emergent phenomenon rather than a fundamental one. There are many curious coincidences in the scientific literature that are somewhat suggestive of this idea, and some of the more noteworthy examples shall be discussed in this subsection. In 1967, Sakharov [72]12 published a short paper in which he demonstrated that a term corresponding to the Einstein–Hilbert action—which leads to Einstein’s field equations by the principle of least action—can be shown to arise from the one-loop interactions of a quantum field theory of interacting fields. In his paper13, Sakharov assumes that there is some background spacetime described by a Pseudo-Riemannian (specifically, Lorentzian) manifold, and that the Lagrangian density describing a quantum field theory atop this manifold can be expressed in the form of a series of powers of the manifold’s scalar curvature ' (the invariant of
the Ricci tensor '). Sakharov allows the background spacetime to be a dynamic object but does not specify any rules governing its dynamical behaviour (that is to say, the Einstein field equations are not explicitly assumed to apply to the spacetime14). Under the aforementioned considerations, the effective action of the fields at the level of one-loop quantum field theory contains terms that are proportional to the cosmological constant ( '0), the Einstein–Hilbert action ∼ ( '), and functions of higher-powers of the scalar curvature ( '2 ). As a result, ∼ O( ) the fields behave as though they exist within a spacetime that obeys the Einstein field
11A term being coined here. This is not, to the author’s knowledge, a term already in use in the literature. 12See reference [73] for an English translation of the original Russian manuscript 13Though not explicitly spelled out—at least, not in the English translation. 14As Visser points out in reference [74], the Lorentzian manifold is “free to flap in the breeze”. The case for not quantizing gravity 15 | equations, despite this particular dynamical description of the manifold never being explicitly assumed. This idea has become known as induced gravity due to the fact that the fields atop the background manifold induce changes to their own action that are mathematically consistent with the terms from general relativity that are responsible for the dynamic nature of spacetime. For a more thorough and informative read, see Visser [74] in which he elaborates on Sakharov’s relatively short paper. Sakharov’s work is not the only example of (apparently) non-gravitational phenomena appearing to mimic—or perhaps lead to—gravitational phenomena. Perhaps the best known example of such a result is contained within a short letter by Jacobson [75], titled “Thermodynamics of Spacetime: The Einstein Equation of State”. Jacobson’s derivation considers the observer-dependent causal horizons that are seen by uniformly accelerating observers (Rindler horizons). As demonstrated by Davies [76] and Unruh [77], these horizons are thermal (this is commonly referred to as the Unruh effect). Under the assumption that Rindler horizons obey an entropy-area15 law analogous to that obeyed by black holes [78, 79], Jacobson demonstrates that heat flux across the horizon leads, via the second law of thermodynamics, to a thermodynamic equation of state that is term-by-term equivalent to the Einstein field equations. The degrees of freedom that constitute the thermodynamic system are the quantum fields close to either side of the causal horizon, and the correlations between vacuum fluctuations in the quantum fields just-inside of and just-outside of the horizon constitute the entropy of the system16. Jacobson’s work, like Sakharov’s, implies that a dynamical description of spacetime and gravitation that is consistent with general relativity can arise from considerations of phenomena that are not explicitly gravitational. In more recent years, Verlinde [80] and Padmanabhan [81–83] have separately shown that gravitational phenomena can emerge from holographic considerations17. Padmanabhan has demonstrated that the dynamical equations governing the evolution of some bulk18 region of a spacetime manifold—that is, Einstein’s field equations—can emerge as a result of thermodynamic considerations at the boundary of that same region of spacetime. In particular, Padmanabhan
15Formally, the area of a black hole is the area of the event horizon. 16The state of the field behind a causal horizon cannot be measured by those observers for which the horizon exists, and so the correlations between fields just inside-of and just-outside of the horizon are not observable to these observers. It is in this sense that the correlations between the fields just-inside of and just-outside of the horizon constitute a measure of entropy to these observers. 17The holographic principle is one such example of these ideas. The holographic principle is a con- jecture born from string theory and one that has links with ideas from black-hole thermodynamics (such as the entropy-area law mentioned above in Jacobson’s work). Padmanabhan stresses in his work that his use of the word is to be understood to be more general than this [83]. 18The term bulk is typically understood to refer to volumes/voluminous regions. 16 Introduction |
demonstrates that regions of spacetime that are in holographic equipartition are static, whereas regions of spacetime that deviate from holographic equipartition undergo time evolution [83] . Verlinde’s work utilizes holographic ideas to demonstrate that one can recover Newton’s gravitational force law as an entropic force [80] by associating the positions of particles in space with information/entropy on an abstract surface (or holographic screen) that bounds a gravitational source. Furthermore, a sketch of a derivation of Einstein’s field equations within the context of the holographic principle is also offered by Verlinde [80]; in this context, general relativity emerges from entropic considerations in the thermodynamic limit. While the field of emergent gravity possesses many curious results, there are serious difficulties faced by the research endeavour as a whole. To this end, see Carlip [84] and the references therein for a fairly comprehensive review of the challenges faced by the emergent gravity research endeavour, with a specific focus on the problems faced by two major classes of emergent gravity model. In addition to the general problems faced by the field as a whole, there are problems with and criticisms of specific models of emergent gravity. In particular, Kobakhidze [85, 86] has pointed out that Verlinde’s particular entropic gravity approach necessarily leads to predictions that disagree with experiments involving ultra-cold neutrons in the Earth’s gravitational field. Additionally, Visser [87] clearly and concisely lists many problems that are inherent to Verlinde’s entropic gravity model, though makes note of the fact that such concerns do not necessarily apply to Jacobson’s and Padmanabhan’s models of emergent gravity on account of the fact that their models differ substantially from Verlinde’s model. More recently, Dai and Stojkovic [88] have pointed out that Verlinde’s model contains errors leading to internal inconsistencies.
1.2.5 Other possibilities: questioning our postulates?
As a closing thought on this subsection, recall from Subsection 1.2.3 the result of Eppley and Hannah [13]. If Eppley and Hannah’s result is indeed correct then the gravitational field being fundamentally classical would lead to violations of so far well-tested physical laws such as, for example, the uncertainty principle [13]. Must such violations be avoided though? The quantum gravity literature itself contains suggestions that the uncertainty principle may require modification close to the Planck length in the form of generalised uncertainty principles (see both [89] and [90] and references therein). If it is indeed the case that quantizing gravity comes at the cost of needing to modify our understanding of the uncertainty principle then it should not be any more troubling for a similar prospect to arise Analogue gravity 17 | from a theory in which gravity is classical (provided, of course, that any such modified uncertainty relation would limit to our current uncertainty principle for the relevant parameter regimes for which we know it should hold true). In attempting to unify our understanding of both gravity and quantum mechanics, we may well find ourselves having to question our most foundational assumptions of physics.
1.3 Analogue gravity
We finally arrive back to where we started at the beginning of this chapter: analogue-gravity models. As was stated at the very beginning, analogue-gravity models are models of non-gravitational physical systems whose behaviour, in certain regimes, can be expressed in the mathematical language of general relativity. Given that models of emergent gravity in particular possess precisely this type of feature, analogue gravity models might prove to be an experimentally accessible method of indirectly testing some of the predictions made by such theories.
1.3.1 Origins
1.3.1.1 Gordon’s optical-metric model
Perhaps the earliest example of an analogue-gravity model—or a potential analogue-gravity model, even if not recognised at the time as such—is the optical- metric model developed by Gordon (of Klein–Gordon fame) in 1923 [91] . The optical-metric model describes an effective metric, , ,19 whose geodesics describe ˆ the path of electromagnetic waves through a curved spacetime—characterized by the physical metric tensor , —filled with a dielectric material whose four-velocity is denoted D . In terms of , , D , and also the dielectric permittivity of free space & and the magnetic permeability of free space , the optical metric tensor is as follows: 1 , = , 1 D D , (1.3.1) ˆ ± − & where Greek indices run over the four spacetime dimensions of general relativity. The choice of plus-or-minus on the right-hand-side is dictated by the metric signature convention of choice: for the , , , signature one picks the positive (− + + +) sign on the right-hand-side, and for the , , , signature one picks the (+ − − −) negative sign on the right-hand-side.
19Note that the circumflex (or hat) is merely used to differentiate the optical metric tensor from the physical metric tensor. The circumflex is not meant to imply that the optical metric tensor is a quantum mechanical operator; the optical-metric model is entirely classical. 18 Introduction |
The physical metric tensor , describes the actual curvature of spacetime, and its corresponding geodesics describe the paths through spacetime followed by non-electromagnetic objects (or objects that do not appreciably interact elec- tromagnetically); for example, the orbits of planets will be, for all intents and
purposes, entirely described by the metric , . However, as mentioned above, it is the geodesics of the optical metric tensor , that dictates the path of electro- ˆ magnetic waves through spacetime. In the particular case of flat spacetime—i.e., when there is no actual, physical spacetime curvature present—the dielectric medium filling space becomes the only source of curvature for the optical metric tensor. If one identifies curvature with gravity, then we find that electromagnetic waves propagating through a flat spacetime that is filled with a dielectric material behave as though they are propagating through a curved spacetime that is not filled with any such material. The optical-metric model in the case of flat spacetime highlights the general feature that characterizes all analogue-gravity models: non-gravitational phenom- ena can be used to simulate gravitational phenomena through the identification of curvature with gravity. In the optical-metric model, this is manifest in the fact that electromagnetic waves (e.g., light) can be made to follow curved paths20 through otherwise flat spacetime via an interaction with a dielectric medium. In all other models of analogue gravity, there will be some analogous physical behaviour: the trajectories of some appropriate object will be expressible in the language of differential geometry (specifically, Lorentzian geometry), and these trajectories will, in general, behave as though they follow curved paths on some background manifold, where the curvature of the manifold is dictated by some non-gravitational physical property of the system.
1.3.1.2 Unruh’s acoustic black-hole model
Gordon’s optical-metric model may technically be—in retrospect—the first ex- ample of an analogue-gravity model, but it is Unruh’s model of an acoustic black hole [1] (or dumb hole21) that seems largely responsible for much of the research in the field of analogue gravity. The importance of Unruh’s model to the analogue gravity community is likely a result of the fact that it provided the first real hope of being able to gain insight—by analogy—into systems for which the phenomenology of quantum gravity is thought to be important, a point that
20Where the notion of a curved path is formalized within the framework of differential geometry. 21Where the word “dumb” here is to be understood to be synonymous with
mute—its original etymology [a]—rather than being a synonym for stupid. Analogue gravity 19 |
Unruh himself makes towards the end of his letter.22 The specific physical system that Unruh considered in formulating the notion of an acoustic black hole was that of a spherically symmetric convergent fluid flow (i.e., a fluid that flows radially inwards towards some singular point), where the fluid in question is barotropic, inviscid, and irrotational (respectively, these mean: the density of the fluid is a function of pressure only, the viscosity of the fluid is equal to zero, and there are no vortices in the fluid flow). If the local fluid flow velocity at some radius ' (defined from the central point that the fluid flow is convergent on) is instantaneously equal to the speed of sound within the fluid itself (taken to be constant), then sound waves originating at or downstream of this characteristic radius will be unable to cross into the region of the fluid upstream of ', as the propagation of sound upstream will be exactly met by the local fluid’s displacement downstream. At the level of a verbal analogy this sounds somewhat reminiscent of a black hole, in which nothing—not even light—can escape from the black hole’s gravitational influence once it has passed a characteristic radius known as the Schwarzschild radius. As it turns out, this analogy goes much further than a verbal one, and this is precisely what Unruh demonstrated: sound waves propagating through this hydrodynamical system can be mathematically modelled as though they are a massless scalar field propagating within the curved spacetime surrounding a black-hole. In this specific hydrodynamical model, the equations of motion that govern the propagation of sound through the fluid in regions close to the characteristic radius ' can be approximated by the null-trajectories (3B2 = 0) of a metric-equation describing a curved geometry. Ignoring the angular components, the metric equation in question is23 2 0 ' 2 2s 2 3B = ( ) 22s A ' 3 3A , (1.3.2) 2s ( − ) − 2 A ' ( − ) where 0 A is the (radius-dependent) background density of the fluid about ( ) which the hydrodynamical equations of motion are linearized, 2s is the speed of sound in the fluid (taken to be constant everywhere), is a small constant with dimensions of inverse-time that is used to linearize the fluid flow velocity in regions close to the characteristic radius ', and is a modified time coordinate (see reference [1] for details). This metric equation description of this system is noteworthy due to the fact that it is term-wise equivalent to the metric equation
22Specifically, Unruh makes the claim, “This system forms an excellent theoretical laboratory where many of the unknown effects that quantum gravity could exert on black-hole evaporation can be modelled.” 23 2 Note that an algebraic error in Unruh’s paper [1] results in a missing factor of 2s in the 3A term. 20 Introduction |
describing the spacetime just outside of the event horizon of a Schwarzschild black-hole. Ignoring the angular components of the metric, the metric-equation of Schwarzschild spacetime is (with the , , , metric signature) given by (+ − − −) 1 ' ' − 3B2 = 1 s 223C2 1 s 3A2, (1.3.3) − A − − A where 2 is the speed of light and 's is the Schwarzschild radius. For radial values
close to 's—e.g., close to the event horizon of a Schwarzschild black-hole—the Schwarzschild metric is approximated by24:
A 's 's 3B2 ( − ) 223C2 3A2. (1.3.4) ≈ 's − A 's ( − ) A direct comparison of Equation 1.3.2 and Equation 1.3.4 reveals that the two metric-equations have the exact same mathematical form! That is to say, the mathematical description of sound propagation in a specific region of this specific hydrodynamical system is governed by an effective metric that is formally equivalent to the metric description of Schwarzschild spacetime in regions close to the Schwarzschild radius, which for a black hole is the region of space close to the event horizon. In the spirit of historical completeness, it should be noted that Visser indepen- dently discovered that certain fluid mechanical systems could be described by the Lorentzian geometry of general relativity [92]. Until Visser’s rediscovery of this fact, Unruh’s work appears to have gone largely unknown within the physics community.
1.3.2 Developments
Gordon’s optical-metric model and Unruh’s acoustic black-hole model are only two specific examples of what is now a whole catalogue of models that are both many and varied. While a full review of the field of analogue gravity and the many models that it contains is outside of the scope of this thesis, a brief overview of some select results from the analogue gravity research endeavour will be included here. For a significantly more detailed overview of the state of the subject as of the year 2011, see the comprehensive review article by Barceló, Visser, and Liberati [93]. Experiments within the field of analogue-gravity have, in some instances, provided us with indirect evidence for the existence of certain physical phenomena
24Unruh expresses the Schwarzschild metric in units where = 2 = 1 in his paper [1]: in these units, the Schwarzschild radius is given by 's = 2". Analogue gravity 21 | that are predicted to arise by general relativity or semiclassical gravity. Of note, the following phenomena have found their first experimental evidence (again, indirectly) within the analogue gravity research endeavour:
1. Hawking radiation has been experimentally observed/inferred in a number of different analogue-gravity models in recent years [94–97]. 2. Cosmological particle production has been experimentally observed/inferred in ion-trap experiments in which the trapping potential is rapidly switched [98, 99]. 3. Superradiance has been experimentally observed/inferred in a hydrody- namical analogue of a black-hole [100].
It is important to re-stress that these experimental results only indirectly provide evidence for the existence of the above physical phenomena: they do not (and, in fact, cannot) directly confirm the existence of the actual phenomena that they emulate.25 The field of analogue gravity has also—be it directly or indirectly—resulted in interesting theoretical developments too. In particular, and as Unruh pointed out in his original publication on acoustic black holes, the greatest utility of analogue-gravity models may be to provide us with a source of insight in regards to understanding problems that arise within the realms of semiclassical gravity and quantum gravity. This optimistic prospect afforded to us by analogue-gravity models is likely the main reason that people are drawn to the field, and a fairly considerable research effort has indeed emerged to this end. Some of the noteworthy theoretical results from (or related to26) the analogue gravity literature include:
1. Hawking radiation and gravity while linked, are not fundamentally linked. Hawking radiation exists in non-gravitational settings, with the funda- mental requirements being the presence of a horizon on some Lorentzian manifold (hence the appearance of Hawking radiation in analogue-gravity models). [104] 2. Planck scale physics might ultimately prove to be an unimportant consider- ation in regards to the process of Hawking emission. This is a fact supported by analogue gravity and analogue-gravity inspired numerical [105–107] and analytical [108–111] studies. These studies imply that the existence of (and the specific mathematical form describing) Hawking radiation is insensitive
25Not only is the evidence indirect, but in some cases it appears to be contentious. See the comment by Leonhardt [101] on the experimental interpretation of reference [96], and also see Steinhauer’s response to Leonhardt’s comment [102]. 26Unruh claims ([103]) that work by Jacobson was inspired by analogue models. 22 Introduction |
to modifications of physics at the Planck scale, easing theoretical concerns that arise over considerations of trans-Planckian modes within the Hawking emission process. In the case of analogue gravity models, the transition from continuum physics to microscopic physics at length scales close to the medium’s discretization length is treated analogously to the expected breakdown of current physical laws at the Planck scale.
The experimental and theoretical results from the analogue gravity research endeavour provide us with some evidence for the idea that it may be possible to view some subset of the phenomenology of general relativity as emergent. Given that modern analogue-gravity models are conceived of to be experimentally accessible in principle, the field appears to be of great utility in probing the phenomenology of emergent gravity theories.
1.4 Making sense of it all: the role of analogue-gravity models in the hunt for a microscopic theory of gravity and the aim of this thesis
Ultimately, the goal of physics is to describe the universe around us. Historically, this has meant that people have observed natural phenomena and have then developed models and theories to explain these phenomena. Often the models and theories that we have constructed have then, in turn, allowed for us to infer the existence of further phenomena. It is by the latter part of this process that we first inferred that their ought to exist a theory of quantum gravity: it appears to be a logical consequence of a universe in which both quantum mechanics and general relativity exist. Subsequently, further results from theory indicated to us that gravity might, in fact, be an emergent phenomena. Whatever the case may be, we have so far found ourselves at a dead-end when it comes to experimentation: owed to the extreme and remote physical circumstances in which the effects of both quantum mechanics and gravity are expected to be important, we have as of yet had no means in obtaining experimental evidence to direct our research efforts. In regards to theory development, there are two major concerns that arise from our lack of ability to perform experiments. First, even if we happened to come across a valid microscopic theory of gravity, an inability to experimentally validate it would prevent us from determining that it is indeed valid. Second, without experimental guidance it is very difficult to put bounds on which parts of theory-space we should explore. Making sense of it all: the role of analogue-gravity models in the hunt for a microscopic theory of gravity and the aim of this thesis 23 | So how does all of this tie together? We want to unify our understanding of gravity with quantum mechanics, and to date we have had no experimental guidance in theory development. We do, however, know of analogue-gravity models: models that mimic some of the phenomenology of general relativity, and importantly, models that are based on physical systems that can actually be (and have actually been) constructed within laboratories. In the absence of any direct tests for our microscopic theories of gravity, indirect tests are the next best thing. Provided that we can identify appropriate candidate systems for a given type of gravitational phenomena, analogue gravity models provide a compelling avenue by which we can indirectly test such phenomena. How best do we go about utilizing analogue-gravity systems to this end then? In some sense, the human experimenters who would deal with an analogue- gravity experiment have a “god’s eye view” on any such analogue gravity universe: they know the rest frame of the medium and they can signal faster (via light) than should be allowed (via sound) in the analogue universe. The actual experimenter’s view—or the laboratory view—is therefore not appropriate if one wishes to use analogue-gravity models as an operational substitute for an actual astrophysical or cosmological experiment. It seems important then to devise some operational procedure by which we can investigate analogue gravity models internally: that is, from the point of view of observers or devices for whom the analogy to relativistic physics appears to be identity, rather than just an analogy. That brings us to the aim of this thesis. At a broad level, the aim of this thesis is conceptually quite straightforward: the aim is to develop and characterize notions of observers/devices/detectors that are natural to analogue gravity universes. By “natural to analogue gravity universes” it is meant that these observers/devices/detectors are constructed and operated in such a way that they obey the symmetries that are inherent to the analogue gravity universe. To this end, the following restriction is placed on the observers/devices/detectors that will be herein considered: any measurements or communication of information can only be made using sound. In the acoustic models of analogue gravity, sound is the object whose trajectories are the null- geodesics of some curved Lorentzian metric, and so by restricting measurements and communication in this way, we can—in the appropriate regimes—consider our observers/devices/detectors to be one-to-one operational substitutes for the observers/devices/detectors that we would ideally have access to in a real (i.e., non-analogue) gravitational experiment.
THEORY
This chapter is intended to provide a somewhat pedagogical overview of necessary mathematical machinery required to understand the field of analogue gravity at the broadest of levels. In essence, this chapter should provide the reader with two things: first, a self-contained reference point to make sense of the notation utilized throughout this thesis; second, the basic conceptual understanding required to make sense of the content of this thesis. 2 The reader who is unfamiliar with—or who feels slightly rusty in regards to—quantum theory and relativistic physics will hopefully benefit from taking the time to read through Sections 2.1 and 2.2; the reader who is reasonably comfortable with these areas of physics can safely skip the corresponding section (or sections) without fear of missing anything of vital importance. Section 2.3 provides some of the basics of field theory—both classical and quantum—and will, as per the preceding sections, be of use to the reader who is either unfamiliar with the field, or the reader who feels that they are in need of a brief recap on the subject matter. Sections 2.4 and 2.5 can be seen as complimentary to one another: both sections detail the construction of specific analogue gravity models, however, while Section 2.4 takes a top-down approach to describe an analogue gravity model based on a particular physical system, Section 2.5 takes a bottom-up approach to describe an analogue gravity system using a toy model. In particular, the former of these two sections—Section 2.4—details the construction of perhaps the most famous class of analogue gravity models: this section is a vital necessity to the reader who is unfamiliar with the details of analogue gravity models; conversely, this section can safely be skipped by the reader with a familiarity on the subject matter. If the reader is interested in specific models of analogue gravity—or the investigation of specific phenomena within analogue gravity models—then a more technical and detailed understanding of the topics discussed within this chapter (and topics not discussed herein) will likely be required. References will be provided throughout for the reader who wishes to read more deeply on the relevant subject matters.
2.1 Relativity
The path that leads to special relativity can be appreciated with a basic observation: from Maxwell’s equations, one can derive a wave equation for electromagnetic waves—of which light is an example—that makes reference to some characteristic
25 26 Theory |
speed: the speed of light. A speed, one would think, should be made with respect to some rest frame; for example, the speed of sound of some given material is defined with respect to the rest frame of the material itself. This leads to the question: with respect to which frame is the speed of light defined? For a not-inconsiderable period of time it was assumed that space—even though it appeared to be a vacuum, totally devoid of anything material—was filled with some light-bearing medium: the luminiferous aether. The conventional wisdom prior to the early 20th century was that it was with respect to the luminiferous aether that the speed of light was defined, though eventually the nature of the luminiferous aether was called into question on both theoretical and experimental grounds: despite repeated attempts to detect the presence—or rather, the effect of the presence—of this hypothetical material, it could not be observed. It is usually considered (at least in retrospect) that the final nail in the coffin for aether theory came after the null result of the famous Michelson–Morley experiment in 1887 [112]: by this point in time, significant developments in optics had permitted for the construction of an interferometer sensitive enough to allow for the detection of the difference in the speed of light that should occur in each arm of the interferometer over the course of a year if the Earth moved through a medium that carried light. No such evidence could be found. There were efforts to save aether theory after the null-result of the Michelson– Morley experiment, and this is discussed somewhat more in Section 3.1, however, this is not our focus here. Our focus here is to give a brief overview of the mathematical machinery of Einstein’s theories of special relativity and general relativity, both of which are constructed under the assumption that there is no preferred universal reference frame.
2.1.1 Tensors and Einstein summation convention
The mathematical foundation of special relativity and general relativity is that of differential geometry, and within the scope of differential geometry, the particular mathematical stage for special relativity and general relativity is that of a pseduo- Riemannian geometry [113, 114]. While a proper review of differential geometry is outside of the scope of this thesis, this subsection will endeavour to provide the reader with the necessary notational rules to understand the content of this thesis (where differential geometry is concerned), and in the process of presenting this some of the basic concepts of differential geometry will be discussed. Before proceeding any further, let us take a moment to make clear the conventions and notational choices that will apply herein:
• We will make use of the coordinate free notation of tensors to refer to Relativity 27 |
tensors in the abstract. In this notation, tensors are denoted with bold font and no indices, e.g., A. When confronted with a symbol like this, one is to understand that we are talking about the tensor A, rather than its individual components or any specific representation thereof.
• In referring to tensors in regards to their components (be it explicitly or implicitly) we will use abstract index notation. Unless otherwise specified, indices do not refer to any particular coordinate system. For example, on its own the object is to be understood to refer to the collection of components that constitute the mixed rank-2 tensor A.
• Primed and non-primed symbols will be utilized to distinguish between representations of the same tensor in different coordinate systems. Given the rank-2 tensor A, the non-primed symbol will be used to denote the components of A with respect to one arbitrary coordinate system (G = G1,G2,G3,G4 ), whereas the primed symbol will be used to denote the ( ) 0 components of A with respect to some other arbitrary coordinate system (G = G 1,G 2,G 3,G 4 ). 0 ( 0 0 0 0 ) In general these conventions and notational choices will apply throughout the rest of this chapter unless otherwise stated. The mathematical equations that govern special relativity and general relativ- ity are expressed in terms of tensorial equations. Tensorial equations are expressed in terms of tensors1, which are mathematical objects that, by construction, possess coordinate-independent properties (that is, each tensor is associated with invari- ant properties under coordinate transformations). The coordinate independent properties of tensors are inherited by tensorial equations, and as a result, these equations are useful for describing physical laws: coordinates are not physical, and therefore, coordinates should play no explicit role in the laws of physics. The non-physicality of coordinates is highlighted in Figure 2.1: two different observers come across some vectorial quantity in space whose existence is entirely separate of the coordinate systems used by those observers to describe it. The vector simply is. The coordinates are an artificial choice made in the minds of the observers. A tensor of rank-# in a -dimensional space is specified by # components or elements. In abstract index notation the rank of a tensor is captured by how many indices must be used to denote its components fully—one index per rank—and the indices themselves take values as many as there are dimensions. For example, a specific rank-2 tensor M in 3-dimensional space can be referred to in terms of its components "8 , where 8, 9 1, 2, 3 : we have two indices for two ranks 9 ∈ { } 1Or, more correctly, tensor fields. 28 Theory |
Figure 2.1: Two observers find a vector floating in space. One observer thinks that a polar coordinate system should be used to describe the vector, but the other observer thinks that a Cartesian coordinate system is the more obvious choice. In reality, the vector’s existence is not contingent on the existence of some coordinate system: it simply is. There is no coordinate system floating around in space with the vector, and the coordinate systems that are imagined by the two observers are merely that: imagined. Coordinate systems are a useful abstract mathematical tool, but unlike the things that they describe, coordinates are not inherently physically meaningful.
(# = 2), each index runs over three values because our space is three-dimensional ( = 3), and so in total we have # = 32 unique combinations of the ordered pair 8, 9 , where a given 8, 9 denotes an element/component of M. ( ) ( ) Tensors of rank-1 and rank-2 can be conveniently represented on paper as arrays. In relativistic physics, we work with = 4 and we usually start our indexing with 0 rather than 1. As a result, in what follows we will most often see tensors denoted in the following way when an explicit matrix representation is given:
0 * ,00 ,01 ,02 ,03 © 1ª © ª * ® ,10 ,11 ,12 ,13® * = ® , = ® . (2.1.1) 2® ® * ® ,20 ,21 ,22 ,23® 3® ® * ,30 ,31 ,32 ,33 « ¬ « ¬ The position of the indices of a tensor—superscript (upstairs) or subscript (down- stairs)—serve two main purposes: it provides both a notational way to keep track of the mathematical structure of a given tensor, as well as a notational shorthand for performing mathematical operations on the components of tensors. The math- ematical structure that is indicated by the index positions is the following: a tensor Relativity 29 | with only superscript (upstairs) indices is referred to as contravariant, whereas a tensor with only subscript (downstairs) indices is referred to as covariant. Tensors do not have to be completely contravariant or completely covariant: they can be mixed and have both contravariant and covariant properties, e.g., the Riemann curvature tensor often appears in mixed form as ' . The notational shorthand (which is itself a consequence of the mathematical structure of tensors) afforded to us by the index placement is that of Einstein summation convention. Einstein summation convention is the following: when an index is repeated once upstairs and once downstairs, it is to be summed over. Two examples to highlight this are:
1 Õ− 0 1 1 B = 0 1 − 1, (2.1.2) + + · · · + − =0 1 Õ− 0 1 1 B = − . (2.1.3) 0 + 1 + · · · + 1 =0 −
In terms of Einstein summation convention, Equation 2.1.2 is an example of the dot product between two vectors, whereas Equation 2.1.3 is an example of matrix multiplication of a vector. Using Einstein summation convention, we can write the components of a vector in some coordinate system (primed) in terms of the components of the same vector in some other coordinate system (non-primed). The transformation rule for contravariant indices is:
%G0 0 = ; (2.1.4) %G whereas covariant indices transform as follows:
%G 0 = . (2.1.5) %G0 A point of notation: note that when an index appears upstairs in the denominator, it functions as though it were downstairs overall (hence the summation over in Equation 2.1.4, leaving only left). The opposite is also true: an index that is downstairs in the denominator functions as though it were upstairs overall. Also note that these transformation rules apply to a given index (rather than “to the tensor”, and so the transformation rule generalizes in a straight-forward manner to tensors of higher rank. For example, the components of the Riemann curvature tensor, ' , will transform like:
%G %G %G %G ' = 0 ' (2.1.6) 0 $ $ %G %G0 %G0 %G0 30 Theory |
Contravariant rank-1 tensors are just vectors—which we shall refer to as contravari- ant vectors—whereas covariant rank-1 tensors are one-forms—which we shall refer to as covariant vectors.
The rank-2 tensor of partial derivatives that transforms the components of covariant vectors between coordinate systems in Equation 2.1.5 is the Jacobian matrix relating those two coordinate systems (denoted J), whereas the rank-2 tensor of partial derivatives that transforms the components of contravariant vectors between coordinate systems in Equation 2.1.4 is the inverse Jacobian matrix 1 (denoted J− ). The explicit matrix representation of the Jacobian matrix and the inverse Jacobian matrix take the following forms:
0 0 %G0 %G0 %G %G © 0 0 ª © 0 0 ª %G ··· %G ® %G0 ··· %G0 ® %G0 . . . ® 1 %G . . . ® = = . . . . ® ; − = = . . . . ® ; %G ® %G ® 0 ® 0 0 ® %G0 %G0 ® %G %G ® ··· « %G ··· %G ¬ «%G0 %G0 ¬ (2.1.7) where as previously established, the indices from 0 to , where is the dimen- sionality of the space. The Jacobian matrix and the inverse Jacobian matrix obey the following identity:
1 %G0 %G 1, = , − = = = (2.1.8) %G %G 0 0, ≠ , where the collection of components constitute the Kronecker delta tensor.
A rank-2 tensor of supreme importance to differential geometry and tensor algebra is the metric tensor g. For our purposes, it is sufficient to say that the metric tensor is defined such that it maps pairs of contravariant vectors to the reals, R in the following way: , = , (2.1.9) where R is a scalar (or a rank-0 tensor). Notationally, this means that the ∈ metric tensor acts on contravariant vectors in such a way as to lower their index (i.e. the metric tensor turns upstairs indices into downstairs indices, turning contravariant vectors into covariant vectors) and swap the label:
, = , (2.1.10) , = , (2.1.11) Relativity 31 | which can be used to induce the summation implied by Einstein summation convention in Equation 2.1.9:
, = = = (2.1.12)
Note that while we may think of defining the metric tensor such that it maps pairs of vectors to the reals, we are not restricted to only use it in this way. The index-lowering property of the metric tensor as seen in Equations 2.1.10 and 2.1.11 is valid on its own, which is to say we can generically lower indices without inducing summation if we so please:
, = . (2.1.13)
The metric tensor can equally well be used to lower the contravariant indices on tensors of arbitrary rank, e.g.,
, , = , . (2.1.14)
From the metric tensor, we can define the inverse metric tensor whose components , satisfy the following propety:
1, = , ,, = = . (2.1.15) 0, ≠ . The inverse metric tensor acts on covariant vectors to raise their index (i.e. it turns covariant vectors into contravariant vectors), in analogy to how the metric tensor acts on contravariant vectors:
, = , (2.1.16) , = . (2.1.17)
By analogy with the metric tensor, the inverse metric tensor can equally well be used to raise the covariant indices of higher order tensors, e.g.,
, , = , . (2.1.18)
A process called tensor contraction—a generalization to the trace of a matrix—can be performed on mixed type tensors with rank-2 or higher. Contraction of a tensor occurs when a contravariant index and a covariant index of that tensor are summed over: the result of this process is to reduce a tensor of rank-# to a 32 Theory |
new tensor of rank- # 2 . For example, the Ricci curvature tensor is formed by ( − ) contracting the Riemann curvature tensor over its contravariant index and its second covariant index:
' B ' = '0 '1 '2 '3 . (2.1.19) 0 + 1 + 2 + 3
When a tensor of rank-2 or greater has more than one contravariant index, the metric tensor can be used to lower one of the contravariant indices and induce contraction. Likewise, when a tensor of rank-2 or greater has more than one covariant index, the inverse metric tensor can be used to raise one of the covariant indices and induce contraction. For example, the inverse metric tensor can be used to contract the Ricci curvature tensor to form the Ricci scalar, a measure of the scalar curvature of a manifold:
' B ,' = ' = ' = ' 0 ' 1 ' 2 ' 3. (2.1.20) 0 + 1 + 2 + 3
The metric tensor—or objects that can be formed from it—captures all of the relevant geometric information of the space that it describes. In a sense, the metric tensor is a full-specification of geometry. One of the most obvious geometric quantities that one often needs to compute in a given geometric space is the length of some path connecting two points. The total length B of some path connecting two points can be computed by integrating the infinitesimal line-element 3B along the path. The infinitesimal line-element, 3B, is given in terms of the coordinate differentials 3G in the following way:
2 3B = , 3G 3G . (2.1.21)
In ordinary Euclidean space described with Cartesian coordinates we have the following
G 1 0 0 © ª © ª 8 ® ® G = H® , ,89 = 0 1 0® , (2.1.22) ® ® I 0 0 1 « ¬ « ¬ from which we can show that Equation 2.1.21 reduces to Pythagoras’ theorem in three dimensions:
3B2 = 3G2 3H2 3I2 (2.1.23) + + Relativity 33 |
2.1.2 Special relativity
Before proceeding, we make some additional notational choices for this subsection:
• Unless otherwise specified, when the matrix representation of a tensor is given, the Cartesian spatial coordinates G,H,I will be assumed. That is to say, ( ) when the symbol appears on one side of an equals sign, and a matrix corresponding to appears on the other side of that same equals sign, the spatial coordinates referred to by and are assumed to be G,H,I by ( ) default. Note that the abstract index notation convention that we established in Subsection 2.1.1 is overruled in this particular case but not in general (i.e., when the symbol appears on its own with no reference to a matrix, abstract index notation still holds). • Any reference frames under consideration are taken to have aligned spatial coordinate axes. That is to say, if G,H,I and G ,H ,I are the spatial ( ) ( 0 0 0) coordinate systems of two distinct reference frames, then G and G0 are
parallel, H and H0 are parallel, and I and I0 are parallel. • When a specific pair of reference frames are being considered we will assume that there exists some point in time at which their spatial origins coincided. Furthermore, we will assume that clocks positioned at the spatial origins of both reference frames read the same time when both spatial origins are coincident. We will define this time to be zero. In other words, both reference frames share a common origin with respect to both spatial and temporal coordinates. • Any relative motion between pairs of reference frames is restricted to the G-direction (as is conventional), and we take the primed frame to be moving with a positive velocity with respect to the non-primed frame. The historical path—that taken by Einstein—to arrive at the theory of special relativity (alternatively: the special theory of relativity or just special relativity) is based on two physical postulates/principles:2 1. The principle of special relativity: The laws of physics should be the same in any inertial reference frame.3 2. The principle of invariant speed of light: The speed of light takes the same value to all observers. 2There are unspoken mathematical assumptions being made here too. For example, it is assumed that the transformation laws that relate the coordinate systems of different reference frames are linear in nature. 3In other words, the equations that govern the fundamental physics of the universe should not be velocity dependent. That is not to say that velocity should not appear in the equations that govern physical phenomena, but rather that the mathematial structure of the equations themselves should not be velocity depenent. 34 Theory |
Special relativity is often described as “putting space and time on equal footing”, in that time becomes a dynamical variable of the theory rather than just a parameter: physical objects are treated as existing within a four-dimensional geometric space referred to as spacetime—a combination of the three spatial dimensions that we see in reality and the one time dimension that we experience—rather than merely existing in a three-dimensional geometric space with time as a parameter. Mathematically, spacetime is treated as a four-dimensional pseudo-Riemannian manifold characterized by a metric tensor4 that is everywhere nondegenerate5, but that—in contrast to the metric tensor describing a Riemannian manifold—allows for proper-lengths that are negative or zero. The spacetime of special relativity is characterized by the Minkowski metric, whose components are most commonly
labelled . In Cartesian coordinates, the Minkowski metric can be given in one of two ways: 1 0 0 0 ©∓ ª ® 0 1 0 0 ® = ± ® . (2.1.24) 0 0 1 0 ® ± ® 0 0 0 1 « ± ¬ In this subsection we will choose to work with the mostly-positive metric signature, , , , 6. That is to say, we will be working with the Metric: (− + + +)
1 0 0 0 ©− ª ® 0 1 0 0® = ® . (2.1.25) 0 0 1 0® ® 0 0 0 1 « ¬ Let’s make our lives easier by pre-emptively defining the fractional velocity (with respect to light) () and the Lorentz factor () as follows:
E B , (2.1.26) 2 1 B p ; (2.1.27) 1 2 −
4We shall often simply refer to the metric describing a spacetime as “the metric tensor” or “the metric”. 5The requirement that the metric tensor be nondegenerate can be given by the following statement: given a non-zero tangent vector U (i.e., the length of U is non-zero) at any point of the manifold, one can always find a tangent vector V at the same point of the manifold such that ,+ * ≠ 0. 6By contrast, the mostly negative metric signature is , , , . The mostly-negative metric (+ − − −) signature will be utilized in other parts of this thesis, and so to prevent confusion the reader will always be made aware of the metric signature of choice in a given scenario. Relativity 35 | where E B v and v is the velocity of an object as is commonly understood. | | Equipped with these definitions, we can now delve in to the formalism of special relativity.
The four-position X can be represented in flat spacetime using Cartesian spatial-coordinates as follows:
2C © ª ! ® G ® 2C - = ® = , (2.1.28) H ® x ® I « ¬ where we have defined the three-position vector x B G,H,I . In special relativity ( ) the time C is merely another coordinate, and in general different observers will associate different values of C to a given event. The coordinates of one reference frame can be related to coordinates of another reference frame in special relativity using the Lorentz transformation, which is given by the rank-2 tensor Λ, and from 1 which we will denote the inverse Lorentz-transformation as either Λ− or Λ0 (the suggestive notation, as we will see, is intentional). The action of these tensors on the four-position is given as follows:
-0 = Λ - , (2.1.29) 1 - = Λ− -0 (2.1.30)
In its most general form, the matrix representation of Λ is quite cumbersome to work with. Thankfully with our particular choices (Cartesian coordinates, both reference frames have aligned coordinate axes and share a common origin, and the primed frame has positive velocity with respect to the non-primed frame), the Lorentz transformation and its inverse can be represented simply as follows:
0 0 0 0 © − ª © ª ® ® 0 0® 1 0 0® Λ = − ® ; Λ− = Λ0 = ® . (2.1.31) 0 0 1 0® 0 0 1 0® ® ® 0 0 0 1 0 0 0 1 « ¬ « ¬
Written out in matrix form this way, the alternate notation of Λ0 is slightly more clear. The only difference between the representations of the Lorentz transformation and its inverse under our assumptions is that the sign on differs. If the non-primed frame measures the primed frame to have fractional velocity , then the primed frame measures the non-primed frame to have fractional velocity 36 Theory |
: in this sense, the definition of the Lorentz transformation matrix in one frame − corresponds to the definition of the inverse Lorentz-transformation matrix in the 1 other frame, hence the notational choice to have Λ− = Λ0. Let us explicitly use Equation 2.1.30 to write the components - in terms of the components -0 :
2C 0 0 2C 2C G © ª © ª © 0ª © 0 + 0 ª ® ® ® ® G ® 0 0® G0 ® G0 2C0 ® ® = ® ® = + ® (2.1.32) H ® 0 0 1 0® H0 ® H0 ® ® ® ® ® I 0 0 0 1 I0 I0 « ¬ « ¬ « ¬ « ¬
While the times C and C0 are coordinates, there does exist a coordinate independent notion of time within special relativity: the proper time between two events in spacetime is the time elapsed by any inertial clock that is coincident with both events. The proper time is unambiguous and is agreed on by all observers: an inertial clock that is coincident with both events will have a definite, observer-independent time on it when it is coincident with the first event, and it will also have a definite, observer-independent time on it when it is coincident with the second event. If, for the sake of illustration, both events are tied to a single physical object (e.g., a point-like light source turning on at one moment in time and then turning off again at some later point in time), then the proper time can simply be understood to be the time ellapsed between those two events by a clock that is attached to that object. Said another way, the proper time elapsed between two events is the difference in the time coordinate in the frame of those events. The mathematical relation between some observer’s coordinate time and the proper time associated to a pair of events can be obtained easily by considering the components of the four-position X in two separate frames. Take a pair of events to define some primed frame in the following sense: the first event occurs at the coordinates C ,G ,H ,I = 0, 0, 0, 0 , and the second event occurs at the ( 0 0 0 0) ( ) coordinates C ,G ,H ,I = C0 , 0, 0, 0 , where the subscript in C0 is used to remind ( 0 0 0 0) ( 2 ) 2 us that this is the value of the coordinate C0 for the second event. Because both events occur at the same spatial coordinates in the primed frame, any clock that is both stationary with respect to the primed frame and coincident with the first event is then also coincident with the second event. As a result, the proper time between both events is then simply given by the difference in the coordinate
times C0 at the location of those events in the primed frame: = C20 . The Lorentz transformation relating the non-primed frame to the primed frame can then be used to show the following: 3C = . (2.1.33) 3 Relativity 37 |
This is one way to express the notion of time dilation. In the non-primed frame time increments forwards in amounts larger than in the primed frame by a factor of .
From here we are properly equipped to define the four-velocity U. The compo- nents of the four velocity can be defined in the following way: ! 3- 3C 3- 3- 2 * B = = = (2.1.34) 3 3 3C 3C v
One can quickly verify that the four-velocity is indeed a proper tensor using the Minkowski metric. The dot product of the four-velocity with itself yields 22, where 2 is the speed of light: in relativistic physics the speed of light is a − constant to all observers, and so the four-velocity has a characteristic invariant to all observers:
** = ** = 2 22 E2 = 222 1 2 = 22, (2.1.35) (− + ) − ( − ) − where the final equality follows from the definition of the Lorentz factor.
Using the definition of the infinitesimal line element from differential geometry (as defined in Equation 2.1.21), one can express the differential spacetime line-element in terms of the differentials of either the four-position or the four-velocity:
2 2 3B = 3 * * = 3- 3- (2.1.36)
3- 3- is straight-forward to evaluate, and * * is given by Equa- tion 2.1.35. Explicitly evaluating the right-most term with Cartesian spatial coordinates gives the following:
3B2 = 223 2 = 223C2 3G2 3H2 3I2. (2.1.37) − − + + +
To actually compute the length of the spacetime line-element B we can write the following: s ¹ B B0 ¹ 0 ¹ C C0 2 2 2 + + + 1 3G 1 3H 1 3I 3B = 82 3 = 82 1 3C (2.1.38) ˜ ˜ − 2 2 − 2 2 − 2 2 ˜ B0 0 C0 2 3C 2 3C 2 3C ˜ ˜ ˜ C C0 ¹ + q = 82 1 C,G,H,I 23C, (2.1.39) ˜ ˜ C0 − ( ) ¹ C C0 + 3C = 82 ˜. (2.1.40) C0 38 Theory |
Note that we have used tildes above symbols here (e.g., C) to denote when a ˜ variable is an integration variable. The square of the spacetime line-element7 can then be expressed like:
!2 ¹ C C0 + 3C B2 = 22 2 = 22 ˜ . (2.1.41) − − C0
The right-most expression is quite useful as it provides observers in any generic frame an explicit method by which to parametrize (and thus compute) the spacetime line-element between two events in terms of the time that they measure between those two events. The sign of B2 between any two events has an important physical meaning within relativistic theories, dividing spacetime into three distinct regions in the following way:
• s2 < 0 characterizes trajectories through spacetime that are traced out by particles with positive rest-mass. These are called time-like trajectories or time-like curves, and pairs of events along such trajectories are referred to as being time-like separated.
• s2 = 0 characterizes trajectories through spacetime that are traced out by massless particles. These are called null trajectories or null curves, and pairs of events along such trajectories are called light-like separated.
• s2 = 0 > 0 characterizes trajectories through spacetime that cannot be traced out by any known particle as, in order for a particle to trace out such a trajectory, it would have to travel faster than light.8 We refer to these paths through spacetime as space-like curves, and pairs of events along such curves are referred to as being space-like separated.
The regions of spacetime corresponding to the three distinct signs of B2 provide us with a notion of causal structure within relativity: events can only be causally related to one another if they are connected by light-like or time-like trajectories, i.e., for two events to be causally connected they must be characterized by B2 0; ≤ if two events are characterized by B2 > 0 then they cannot possibly have causally influenced one another (because no real particle can possibly have travelled such a path).
7This is merely convention; we could equally as well write out an equation for B itself, however it is customary in relativistic physics to write down the square on account of the fact that B2 is always a real number (B2 R). ∈ 8Hypothetical particles called tachyons would be able to trace out such trajectories if they existed, though there is no evidence for the existence of such particles. Relativity 39 |
2.1.3 General relativity
For a comprehensive guide to Einstein’s general theory of relativity, consult Gravitation9 by Misner, Thorne and Wheeler [113]. Once special relativity has been established, the principle that leads to general relativity is the the equivalence principle. Conceptually, the principle takes the following form:
• The equivalence principle: the inertial mass of an object is equivalent to its gravitational mass. Equipped with the equivalence principle, one can proceed to construct a theory of gravity within the mathematical framework of differential geometry. The spacetime of special relativity is described by a metric tensor field that has— to inertial observers—the same description everywhere and at all times: the
Minkowski metric, whose components are denoted . On the contrary, the metric tensor field in general relativity—whose components are denoted ,—is much more generic: in general, it is a function of both space and time. In general relativity, we identify the curvature of spacetime—how the spacetime manifold (described by the metric tensor) varies from point-to-point in space and time—to be what we refer to as gravity. Furthermore, unlike in special relativity in which the metric tensor field forms a static background upon which events simply happen to unfold, the metric tensor field of general relativity is a dynamical variable whose structure is determined by the matter and energy content of spacetime itself. The dynamical equations that determines the spacetime metric are called the Einstein Field Equations (EFEs):
1 8 ' ', Λ, = ) (2.1.42) − 2 + 24 where ' are the components of the Ricci curvature tensor, ' is the Ricci scalar
(or the scalar curvature) obtained by contracting the Ricci curvature tensor, , are the components of the metric tensor itself, Λ is the so-called cosmological constant10 (usually identified to be the intrinsic energy of empty space), is
Newton’s gravitational constant, 2 is the speed of light in vacuum, and ) are the components of the stress-energy tensor (or the energy-momentum tensor), which captures the distribution of matter and energy in spacetime. The only non-scalar quantity in this equation that is not inherently related to the metric tensor is the stress-energy tensor. The Ricci curvature tensor (and, by extension, the Ricci 9Be warned though: owed to its truly comprehensive nature on the subject matter, this book is immense (both in terms of scope of content and physical size). 10Note that the cosmological constant Λ is a scalar that is in no way associated to the rank-2 Lorentz transformation tensor Λ. 40 Theory |
curvature ') is given in terms of Christoffel symbols, and Christoffel symbols are themselves given in terms of the metric tensor: as a result, the left-hand-side of the Einstein field equations can be written entirely in terms of the metric tensor and its derivatives. The Einstein field equations therefore describe a dynamic interplay of two conceptually distinct objects: spacetime (represented by the metric tensor), and the matter and energy content of spacetime (represented by the stress-energy tensor). This dynamical interplay is perhaps most famously summed up by John Archibald Wheeler, who said [115]: “Spacetime tells matter how to move; matter tells spacetime how to curve.”
The components of the Ricci curvature tensor, ', are obtained by contracting
the first and third indices of the Rienmann curvature tensor, ' . The Rienmann curvature tensor is defined in terms of the Levi-Civita connection , which itself ∇ can be defined in terms of Christoffel symbols (of the second kind), the components of which are denoted Γ . Using index notation, the Rienmann curvature tensor has components given by
' = % Γ % Γ Γ Γ Γ Γ , (2.1.43) − + −
and so the Ricci curvature tensor’s components can be computed in the following way: ' = ' = , , ' = ' . (2.1.44) where we have relabelled the dummy-index between the left-hand-side and the right-hand-side of the second equal sign. From here, we can further contract the Ricci curvature tensor to obtain the Ricci scalar (or the scalar curvature):
' = ' = , ' = , ' , (2.1.45) where, again, we have relabelled the dummy-index between the left-hand-side and the right-hand-side of the second equal sign. Christoffel symbols are tensor-like objects (Equation 2.1.51 and the preceding discussion elaborate on this point) that it is useful to define in differential geometry, specifically when considering the Levi-Civita connection on a manifold. We define Christoffel symbols of the second kind to be the array of coefficients Γ that can be used to express the Levi-Civita connection in the coordinate direction e like: ∇
e = Γ e . (2.1.46) ∇
Specifically e defines, at every point of a manifold, a local coordinate basis, i.e., it defines the basis of the tangent space at every point on a manifold. The elements Relativity 41 |
of e are themselves rank-1 tensors, rather than scalars, and the form of e can be given by its action on D, which are the components of the tangent vector u:
u = D e. (2.1.47)
Again it is important to stress that the elements of e are not scalars, which can be appreciated by the fact that the action of e on the components of u (that is, the action of e on D ) is to return the vector u itself. In a sense e provides at any point of a manifold the most natural choice of local coordinate basis.
Explicitly, Christoffel symbols of the second kind are given by the equation ! 1 %, %, %, Γ = , . (2.1.48) 2 %G + %G − %G
Perhaps somewhat confusingly, Christoffel symbols of the first kind, Γ, are typically defined in terms of those of the second kind by the equation
Γ = ,Γ . (2.1.49)
Explicitly, Christoffel symbols of the first kind, Γ, are given by the equation: ! 1 %, %, %, Γ = . (2.1.50) 2 %G + %G − %G
While Christoffel symbols are related to tensorial quantities and are expressed in terms of tensorial index notation, it is important to note that Christoffel symbols do not transform like tensors under a change of coordinates. To highlight the point, considering changing from one coordinate system in which the coordinates are labelled with the indices , , and to the coordinate system in which the coordinates are labelled with the indices , , and . A tensorial object should be described by components that transform as per the rules laid out in Subsection 2.1.1—specifically contravariant indices should transform as per Equation 2.1.4, and covariant indices should transform as per Equation 2.1.5— however, a Christoffel symbol obeys the following transformation rule instead:
2 %G0 %G %G % G %G0 Γ0 = Γ (2.1.51) %G %G0 %G0 + %G0 %G0 %G Despite not being tensors themselves, Christoffel symbols play an important role in general relativity. For example, Christoffel symbols are used in the geodesic equation. Geodesics are a notion from differential geometry that are of great physical 42 Theory | importance to general relativity. In differential geometry, a geodesic between two points on a curved manifold is the path of shortest distance between those two points. In general relativity, geodesics are the trajectories that objects naturally follow through spacetime when free of any non-gravitational forces: we call such trajectories the world-line of the particle. The geodesic equation can be given in terms of the proper time as follows:
32G 3G 3G Γ = 0. (2.1.52) 3 2 + 3 3 The world-line of an object through spacetime is the path through spacetime described by the four-position G such that the geodesic equation is obeyed, and the time measured by an object travelling such a path is the proper time .
2.2 Quantum physics
For comprehensive modern treatments of modern quantum mechanics, con- sult the well known textbooks of Sakurai [116] and Townsend [117]. Griffiths undergraduate level quantum mechanics textbook [118] provides an accessible introduction to the topic.
2.2.1 Basics of quantum physics
The mathematical language of quantum mechanics is that of linear algebra. Just as much of classical mechanics can be formulated in terms of vectors (e.g. position and momenta), in quantum mechanics we also use vectors to describe physical systems. In quantum mechanics the vectors that we deal with do not belong to the three-dimensional geometric space of Newtonian mechanics nor the four-dimensional geometric space of relativistic theories—they belong to a more general vector space called a Hilbert space11.A Hilbert space is a type of abstract vector space of arbitrary dimension, specifically a real or complex inner product space that is also a complete metric space. In quantum theories, we use Hilbert spaces that are complex inner product spaces, and the dimensionality of the Hilbert spaces—and thus the vectors—that we encounter range from finite and rather small (two-dimensional Hilbert spaces, for example) to uncountably infinite. The dimensionality of the Hilbert space that one considers when investigating some physical phenomenon is determined by the system itself; exactly how will be elucidated in due course.
11Note, however, that Euclidean geometry is itself an example of a Hilbert space. Euclidean geometry is an example of a finite dimensional ( = 3) Hilbert space that is a real inner product space rather than a complex inner product space Quantum physics 43 |
The state of a physical system in quantum mechanics is encoded in the state vector12; these vectors are elements of the Hilbert space and are, in general, complex-valued (hence the complex part of complex inner product space). In the case that a system is described by a finite dimensional Hilbert space, state vectors can be represented on paper in the intuitive way—as a list of numbers13—however, when the system that is being investigated is described by a Hilbert space whose dimension is uncountably infinite, this notational convenience no longer applies. State vectors are represented by some symbol such as #14 within a ket, i.e., like # . While kets represent vectors that are elements of the Hilbert space, | i bras represent elements of a dual vector space (dual to the Hilbert space) and are represented like # . bra–ket notation shall be used throughout this thesis. h | Physical processes—not only interactions, but also the natural evolution of a free system—are encoded into operators in quantum mechanics: we use a “hat”15 to denote such objects, for example ˆ . Operators act on states to produce other states: when acting on kets, operators act from the left, and when acting on bras, operators act from the right, i.e.,
#new = $ # , (2.2.1) | i ˆ | oldi #new = # $ (2.2.2) h | h old| ˆ
Observables are self-adjoint linear operators corresponding to physical processes that can actually be observed (or measured). Consider some physical system that is associated to some observable described by the operator ˆ. If the system is in the state # then the expectation value of the observable is denoted and is | i ˆ h ˆi given by: = # # . (2.2.3) h ˆi h | ˆ| i States are expressed as a linear combination of basis elements, and a basis is nothing more than a choice of coordinates. Because coordinates are not physically meaningful, we have some freedom in picking a basis to describe our Hilbert space, the corresponding states, and the operators that act on those states (provided,
12Really, there is an equivalence class of vectors that define the physical state: any state vector multiplied by any complex number, 2 C, is another valid state vector. As a result, we should ∈ really say that states are defined by rays rather than vectors, however the convention is to refer to state vectors and we shall not break from this convention. 13Of course, when the Hilbert space’s dimensionality is countably infinite, one can also represent parts of the state vector on paper as a list of numbers too. 14The symbol # is usually reserved specifically for the position-space wavefunction in quantum mechanics and spinor fields in quantum field theory. We will use it as a generic label for any type of state, however if the state that we pick happens to coincides with one of these typical uses, we will explicitly make note of this. 15A circumflex. 44 Theory |
of course, that the basis of choice meaningfully describes the physical system of interest). A particularly useful basis when investigating a particular problem in quantum physics is the eigenbasis corresponding to some relevant observable.
The eigenbasis is the set of states #8 (where 8 is just an index—possibly { | i } continuously valued—to indicate which state we are talking about in the set) such that when the observable ˆ acts on a state from this set we obtain
#8 = 08 #8 , (2.2.4) ˆ | i | i where 08 R. The individual elements of the eigenbasis, #8 , are the eigenvectors ∈ | i of the operator ˆ, while the corresponding real numbers 08 are the eigenvalues of the eigenvector. When the system is in an eigenstate then the expectation value of the observable described by ˆ is simply:
= #8 #8 , (2.2.5) h ˆi h | ˆ| i = #8 08 #8 (2.2.6) h | | i = 08 #8 #8 (2.2.7) h | i = 08 . (2.2.8)
In quantum mechanics, the Hamiltonian operator (or simply, the Hamiltonian) is the generator of time translation. This is encoded in the time-dependent Schrödinger equation: 3 8~ # = # . (2.2.9) 3C | i ˆ | i For certain states called stationary states the so-called time-independent Schrödinger equation applies, # = # , (2.2.10) ˆ | i | i where is the energy of the system when its state is the stationary state # . | i Comparing Equation 2.2.10 with Equation 2.2.4 shows that stationary states are eigenvectors of the Hamiltonian. In reality, stationary states actually do have a time-dependent part however the time-dependency is only present in a complex phase factor, and this does not manifest in any observable way. After some time C, an initial stationary state # 0 will have evolved under the Hamiltonian into the | ( )i state # C given by: | ( )i 8C ~ # C = 4− / # 0 , (2.2.11) | ( )i | ( )i where the complex phase factor is the exponential term in front of the state. We mentioned earlier that quantum systems can be described by Hilbert spaces that range from finite to uncountably infinite in dimensionality. How do we know Quantum physics 45 | the dimensionality of the Hilbert space for a generic system? The answer to this question is the same as the answer to the question, “how many values do I need to specify in order to fully describe the state of the quantum system of interest?” Perhaps the best way to highlight this is by considering one of the most useful models in quantum mechanics: that of the quantum harmonic oscillator.
2.2.2 The quantum harmonic oscillator
The harmonic oscillator is a model of fundamental importance to classical physics, and so it should probably come as no surprise that the quantum analogue to this—the quantum harmonic oscillator—is a system of fundamental importance in quantum physics. Perhaps most obviously, such a model can be used to model the motion of an obviously oscillating quantum system: for example, the motion of diatomic molecules. The usefulness of this model, as we shall see, far exceeds this rather humble application of it. The quantum harmonic oscillator is described, as all quantum systems are, by a Hamiltonian. The Hamiltonian for the one-dimensional quantum harmonic oscillator can be given by ?2 1 = ˆ <$2G, (2.2.12) ˆ 2< + 2 ˆ where ? is the momentum operator and G is the position operator. The position ˆ ˆ operator and the momentum operator obey the following commutation relation:
G, ? B G? ?G, (2.2.13) [ ˆ ˆ] ˆ ˆ − ˆ ˆ = 8~. (2.2.14)
In the position basis the momentum operator has the following representation in one-dimension: % ? = 8~ . (2.2.15) ˆ − %G Using the position basis representation of ?, we can rewrite the Hamiltonian for ˆ the quantum hamonic oscillator in the following way:
~2 32 1 = <$2G (2.2.16) ˆ −2< 3G2 + 2 Recall that the answer to the question, “what is the dimension of our Hilbert space?”, is the same as the answer to the question, “how many values do I need to specify in order to fully describe the state of the quantum system of interest?”. In the position basis, the state of our system can be given by its wavefunction, denoted with the bare symbol #: this represents the quantum amplitude of our system as a function of 46 Theory |
spatial location. In quantum mechanics, we treat space as being continuous, i.e., there are an uncountable infinity of points that make up space, and so to fully specify our wavefunction # we must specify the amplitude for the system at an uncountable infinity of points. The dimensionality of the Hilbert space is given by the dimensionality of the state vectors describing our system and so the Hilbert space of interest is uncountably infinite in dimension. How do we represent our state vector when their dimensionality is uncountably infinite? We simply parametrize them in terms of position, just as we would a continuous function of space: in our case, the wavefunction # has some value for every position. Consider now the eigenstates of the Hamiltonian for the quantum har- monic oscillator as described above. The eigenstates solve the time-independent Schrödinger equation, so we can write: ! ~2 32 1 <$2G # = #. (2.2.17) −2< 3G2 + 2
The family of eigenvectors of this Hamiltonian can be given in terms of Hermite
polynomials, = H . The eigenvectors are represented as continuous functions of ( ) space: 1 4 r ! / 1 <$ <$G2 2~ <$ #= G = 4− / = G , (2.2.18) ( ) √2= =! ~ ~ where = = 0, 1, 2,... , i.e., = N0, and the physicists’ Hermite polynomials are given ∈ by: = = H2 3 H2 = H = 1 4 4− . (2.2.19) ( ) (− ) 3H= Note that here we have defined the physicists’ Hermite polynomials in terms of a continuous variable H and that in our Hamiltonian we have H p <$ G. The = ~ eigenvalues are also given in terms of the variable = and are:
1 = = ~$ = . (2.2.20) + 2
Despite not appearing to exist in the original description of the quantum harmonic oscillator’s Hamiltonian, the quantity = appears to be necessary to completely characterize the system as it must be introduced to solve for the eigenvectors and the eigenvalues of the Hamiltonian. The presence of this variable = means that for every frequency of oscillator we don’t just have a single eigenvector and eigenvalue, but rather an entire family of eigenvectors and eigenvalues—one
eigenvector and corresponding eigenvalue for every = N0. For a given oscillator ∈ (i.e., for some fixed value of $) there is a constant energy gap between adjacent Quantum physics 47 | solutions in its family of solutions (i.e., for solutions described by = and = 1): +
Δ B = 1 = , (2.2.21) + − 1 1 = ~$ = 1 ~$ = , (2.2.22) ( + ) + 2 − + 2 = ~$. (2.2.23)
As a result of this behaviour, a given pair of orbitals in an atom can be modelled as a restricted quantum harmonic oscillator (one for which = only has two values): an electron must absorb or emit a fixed quanta of energy—a single photon of energy ~$—in order to transition between the two orbitals.
Let us now define the following new operators in terms of the position operator, G, and the momentum operator, ?: r <$ 8? 0 = G ˆ , (2.2.24) ˆ 2~ ˆ + <$ r <$ 8? 0† = G ˆ . (2.2.25) ˆ 2~ ˆ − <$
We can show that these operators obey the following commutation relation,
8 0, 0† = − G, ? = 1. (2.2.26) [ˆ ˆ ] ~ [ ˆ ˆ] With these definitions, we can rewrite the Hamiltonian for the quantum harmonic oscillator like: 1 = ~$ 0†0 . (2.2.27) ˆ ˆ ˆ + 2 This particular form of the Hamiltonian is very suggestive; it is reminiscent of the expression for the eigenvalues given in Equation 2.2.20! Taking some inspiration from the eigenvalue equation, define the following operator:
# B 0†0. (2.2.28) ˆ ˆ ˆ
We can also show that this new operator, #ˆ , obeys the following commutation relationships:
#, 0† = 0†, (2.2.29) [ ˆ ˆ ] ˆ #, 0 = 0. (2.2.30) [ ˆ ˆ] −ˆ 48 Theory |
Rewriting the Hamiltonian in terms of the operator # we obtain: | i 1 = ~$ # . (2.2.31) ˆ ˆ + 2
This form of the Hamiltonian is linear in the operator #ˆ and has no other operator dependency, thus the eigenvectors of #ˆ are also eigenvectors of the Hamiltonian. Recall that the eigenvectors of the Hamiltonian obey the time-independent Schrödinger equation: denoting the eigenvectors of # (and thus ) as = , the ˆ ˆ | i time-independent Schrödinger equation can be written as
= = = = , (2.2.32) ˆ | i | i 1 1 ~$ # = = ~$ = = , (2.2.33) ˆ + 2 | i + 2 | i which leads to # = = = = . (2.2.34) ˆ | i | i Because the eigenvalues of the operator #ˆ correspond to how many quanta of energy are in our system, we refer to #ˆ as the number operator, and we call the basis formed by the set of eigenvectors = the number basis (or alternatively the | i Fock basis). We call the individual eigenvectors forming the number basis number states. Note that the number basis is the eigenbasis of the quantum hamrmonic oscillator.
If we act the operator # on the states 0 = and 0 = , we can utilize the ˆ ˆ† | i ˆ | i definition of #ˆ from Equation 2.2.28 and the commutation relations given by Equations 2.2.26, 2.2.29, and 2.2.30 to show the following:
#0† = = = 1 0† = , (2.2.35) ˆ ˆ | i ( + )ˆ | i #0 = = = 1 0 = . (2.2.36) ˆ ˆ | i ( − )ˆ | i
Recall that the operator # tells us that the eigenvalues of the state = are = quanta ˆ | i of energy. We can then determine that the eigenvalues of the state 0 = are = 1 ˆ† | i + quanta of energy, whereas the eigenvalues of the state 0 = are = 1 quanta of ˆ | i − energy. The action of 0 and 0 on = is then to raise and lower the energy of our ˆ† ˆ | i system by a single quanta of energy, respectively. For this reasons, we call the operator 0 the raising operator and we call the operator 0 the lowering operator. ˆ† ˆ Quantum physics 49 |
2.2.3 Scattering theory
A subject that we will need to consider in due course is that of scattering theory. Scattering experiments are one of the best ways to understand a quantum mechanical system, and we will make explicit use of the formalism of scattering theory in Chapter4 . For everything that follows we work in the interaction picture in which both states and operators are time-dependent. Consider a scattering experiment for which we have an incoming quantum system, e.g., some incoming particle or particles, and this incoming quantum system is incident on some target system (possibly quantum, possibly not).
We describe the incoming state of the particle with an in-state, #in , and the | i outgoing state of the particle with an out-state, #out . Assume that the duration | i of interaction16 between the incoming system and the target system is very small compared with both the time of approach (the time between measuring the in-state and the interaction) and the time of flight afterwards (the time between the interaction and measuring the out-state). In reality the interaction may be infinite in range (for example, the electromagnetic interaction) and thus infinite in duration however, when the incoming system and the target system are sufficiently separated, the interaction will be negligibly small. If the incoming quantum system can be described with first quantization (quantum mechanics), and if we insist on modelling the interaction between the target system and the particle in the way that we understand, i.e., with an operator acting on a state, then we can say the following:
#out = ( #in . (2.2.37) | i ˆ | i
That is, the output state is a function of evolving the input state under the action of some operator. What is the operator (ˆ describing a scattering experiment that takes the input state #in and produces the output state #out ? The specifics | i | i depend on the kind of system that we are considering, however a particular useful case to consider is that of weak scattering, i.e., when the scattering can be treated as a perturbation to the system. In this case of weak scattering, we can write the Hamiltonian that governs the time-evolution of the system as a perturbation series in the coupling parameter , which if we truncate to fist order in the coupling parameter is:
C = 0 int C . (2.2.38) ˆ ( ) ˆ + ˆ ( )
Written this way, the time dependency of the Hamiltonian in the interaction
16Somewhat of an ill-defined notion. 50 Theory |
picture has shunted into the interaction term; when the interaction is sufficiently weak, the Hamiltonian is time-independent.
Under the action of this Hamiltonian, the solution to the Schrödinger equation
for time evolution of some initial state at time C0 to some other state at time C 5 can be given by the Dyson series:
= 8 = ¹ C = Õ∞ − Ö Ö 8 8 ~ 0C: 0C: # C = © 3C ª 4 ~ ˆ int4− ~ ˆ # C0 , (2.2.39) | ( 5 )i =! :® T{ ˆ } | ( )i ==0 :=1 C0 :=1 « ¬ where = the desired order of perturbation, and the time-ordering operator17. T If one collects the operator parts of Equation 2.2.39 up into an operator (ˆ – the S-matrix – then Equation 2.2.39 can be simplified to
# C = ( # C0 , (2.2.40) | ( 5 )i ˆ | ( )i which is of the form desired as per Equation 2.2.37. If we express ( as ( = I 8) ˆ ˆ + ˆ then all of the dynamics relevant to the actual scattering process (i.e. those parts of the dynamics that have nothing to do with the initial conditions of the incoming particle) are put into the T-matrix, )ˆ, while the identity accounts for the cases in which the incoming particle fails to scatter from the particle.
Denote our initial state at time C0 to be #in and denote the state that we | i wish to detect after scattering as #out . To characterize scattering, our job is | i then to determine the amplitude associated with the overlap of the actual final
state # C with the state that we desire #out , or equivalently, the amplitude | ( 5 )i | i associated with the initial state #in evolving into the desired state #out : | i | i
#out # C = #out ( #in . (2.2.41) h | ( 5 )i h | ˆ| i
We define a new object denoted to describe the dynamical (rather than ℳ kinematical) aspects of scattering processes: the amplitude for a given state 0 to
transition to another given state 1 defines the matrix element 0 1. In reality ℳ → ℳ can only really be considered a matrix if the dimensionality of the Hilbert space under consideration is finite in size, though the nomenclature sticks regardless of the dimensionality of the Hilbert space. With the relevant elements of the
17The time-ordering operator isn’t a quantum mechanical operator: it’s purpose is to tell you to organize your operators so that everything is ordered in time, i.e. earlier time operators apply to a state before later time operators. Field theory 51 | scattering matrix at our disposal we can then write the following:
4 3 #out 8) #in = 8 in out 2 ( ) ?in ?out in out , (2.2.42) h | ˆ | i ℳ → ( ) ì − ì ( − ) where in out accounts for all of the dynamics of our system (i.e., what interaction ℳ → is actually happening) while the delta functions impose kinematic constraints on the system (i.e., restrict us to the relevant area of phase space).
To obtain a particular description of in out one must have a particular ℳ → system in mind. We will forgo discussing scattering theory in any more depth at this point: additional discussions regarding scattering theory will appear in Section4 when we consider scattering within a particular system.
2.3 Field theory
Field theory is of immense importance in physics in general, and especially so in modern physics: quantum field theory and general relativity—our two best theories of nature to date—are both built atop of field theoretic considerations. Here we shall describe some of the important mathematical waypoints in field theory. At a very basic linguistic level, a physical field is represented by a mathematical object that takes as inputs spatial location and time, and gives in return a single output. The type of output depends on the physical system being modelled: in classical non-relativistic physics scalar fields and vector fields are common; in relativistic physics tensor fields of higher rank are common; in quantum theory one even encounters more exotic fields in the form of spinor fields. We can describe a whole host of spatially dependent quantities in terms of fields: familiar examples are temperature, which takes as input spatial location and time and outputs a real number, and the classical gravitational field from Newtonian mechanics, which takes as input spatial location and time and outputs a vector. While outside of field theory our aim is often to determine trajectories—for example, the position as a function of time, G C , of a ball that is thrown through ì( ) the air—in a field theory, our goal is more ambitious: we want to know the state of the entire field ) G,C (often referred to as the field configuration), which (ì ) usually means the value of the field at every location for a given time or times. In classical or semi-classical physics, very often the state of the field is then itself used to compute the trajectory of some non-field object that interacts with the field: for example, the trajectory of a charged particle through a magnetic field is computed using the state of the field, while the trajectory of an asteroid through 52 Theory |
the solar system can be computed from the state of the gravitational fields of the sun, nearby planets, and other massive bodies.
How do we determine the entire state of a field at some time? In certain cases, the state of the field is easy to compute (or at least easy compared with the general case): if we can compute the state of the field for one of these specific cases then we can determine the state at some future time using the equations of motion of the field. Generally, the equations of motion of fields are some form of wave equation. For example, the three-dimensional linear wave-equation governing some scalar field variable ) is 1 %2) 0 = 2), (2.3.1) −E2 %C2 + ∇ where E is the propagation speed of the waves. There are, of course, more complicated wave equations: this is just the most basic. Plane waves form a basis of solutions to this equation. In general a plane wave is given by
8 $C k x ) x,C = 4− ( − · ) (2.3.2) ( ) where $ is the (angular) frequency of the wave, : is the (angular) wave-vector, and the relation between $ and : is related by the propagation speed E:
E = $: (2.3.3)
That Equation 2.3.2 is a solution to Equation 2.3.1 can be quickly demonstrated:
1 %2) 1 % %) = = :2), (2.3.4) −E2 %C2 −E2 %C %C % %) % %) % %) 2) = = :2 :2 :2 ) = :2), ∇ %G %G + %H %H + %I %I − G + H + I − (2.3.5) 1 %2) 2) = :2) :2) = 0 (2.3.6) −E2 %C2 + ∇ + (− ) As it turns out, the three-dimensional linear wave-equation—Equation 2.3.1— has the same mathematical form as the relativistic wave-equation for a massless particle (where in the relativistic case we have E = 2). Consider, by comparison, the equations of motion for the electric field E and the magnetic field B:
1 %2E 0 = 2E, (2.3.7) − 22 %C2 + ∇ 1 %2B 0 = 2B, (2.3.8) − 22 %C2 + ∇ Field theory 53 | where 2 is the speed of light18. While the fields E and B are not themselves Lorentz covariant objects (though they do originate from a Lorentz-covariant tensor), the overall structure of the wave equation that describes them is invariant under a Lorentz transformation. Any other field whose equations of motion are also given by the linear wave-equation, Equation 2.3.1, with waves travelling at the speed of light, E = 2, will also obey a Lorentz transformation and thus be a relativistic field.
2.3.1 Quantum field theory
A successful path to building quantum models—both first quantized (non- relativistic quantum mechanics) and second19 quantized (quantum field theory)— is by the process of quantization.20 The guide to building a second quantized theory using the process of quantization can be found in many text books, with the process as given here being an adaptation of that from [119]. At the most general level, the path that takes to construct a quantum field theory is the following:
1. Write down a relativistically invariant Lagrangian density for some classical field theory.21 The Lagrangian could describe a single field or several; it could describe a free theory, or a coupled theory. 2. From the Lagrangian density do the following: a) Utilize the Euler–Lagrange equations to find the momentum density of each field in your Lagrangian density; b) Obtain the Hamiltonian density by performing a Legendre transforma- tion on the Lagrangian density. 3. Declare the fields and their associated momentum-densities to be operators that obey commutation/anti-commutation relations for bosons/fermions respectively. 4. Express the field operators in terms of creation and annihilation operators that act on particle number states.
18 The prefactor in front of the time-derivatives is 00—where 0 is the magnetic permeability of free-space, and 0 is the electric permittivity of free-space—when one derives these equations from Maxwell’s equations: this prefactor has units of inverse velocity, and if we identify this as being the square of the speed of light, we obtain the equations in terms of 2. 19The term “second quantization” is an unfortunate vestige that has been carried into the present from the early history of quantum theory. 20The process of first quantization is somewhat mathematically dubious, however. 21Note that the classical field theory that informs your Lagrangian at this step does not need to correspond to some physical system that is actually manifest in the classical world: the only requirement is that you have a Lagrangian that obeys the mathematical structure of a classical field theory. 54 Theory |
To lead by example, let us demonstrate how one can build the most simple kind of quantum field theory: that being the field theory of a free quantum scalar field in Minkowski (flat) spacetime.
1. Write down a classical relativistically-invariant field-theoretic Lagrangian density for the system of your choice.
The form of the simplest relativistic Lagrangian density describing a free scalar- field in Minkowski spacetime is often given as follows:
2 2 1 1 < 2 2 = %)% ) ) . (2.3.9) ℒ 2 − 2 ~2 Note that, written this way, the field ) must have units of square-root-force in order for the Lagrangian density to have the correct units (energy density, or alternatively, pressure). In order to allow ) to have more sane units, we can write the Lagrangian density with a prefactor like:
2 2 ! 1 < 2 2 = %)% ) ) . (2.3.10) ℒ 2 − ~2
If ) has units of distance (for example, if ) describes the displacement of some object away from some equilibrium position) then —as the suggestive notation may imply—has units of pressure. The prefactor of 1 2 become useful for our / next step.
2a. From the Lagrangian density: Utilize the Euler–Lagrange equations to find the momentum density of the field.
The Euler–Lagrange equation for a relativistic field is:
% © % ª ℒ % ℒ ® = 0, (2.3.11) %) − ® % %) « ¬ the momentum density, Π G , of the field ) at some location G is the parenthetical ( ) term in the Euler–Lagrange equation:
% Π G B ℒ . (2.3.12) ( ) % %) G ( ) Field theory 55 |
In our case, the momentum density for the Lagrangian density specified in Equation 2.3.10 is