DYNAMICS OF 3D SUSY COULOMB BRANCHES
A Dissertation Presented to the Faculty of the Graduate School of Cornell University
in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
by Mario Claudio Martone January 2015 c 2015 Mario Claudio Martone
ALL RIGHTS RESERVED DYNAMICS OF 3D SUSY COULOMB BRANCHES Mario Claudio Martone, Ph.D. Cornell University 2015
In this thesis we focus on the dynamics of Coulomb branches in 3 dimensional N = 2 SUSY theories. The first part of this thesis is dedicated for the most part to a review of previously known results, providing the necessary background to understand the generalization to more complicated theories. We review in details both the behavior of U(1) theories with generic number of flavors and SU(N) theories with fundamentals. We also provide insights on the mapping of Coulomb branch operators across dualities as vortex/monopole operators. In the second part of the thesis we investigate the IR dynamics of N = 2 SUSY gauge theories in 3D with antisymmetric matter. The presence of the antisymmetric fields leads to further splitting of the Coulomb branch. Counting zero modes in the instanton background suggests that more than a single direction along the Coulomb branch may remain unlifted. We examine the case of SU(4) with one or two antisymmetric fields and various flavors in detail. Using the results for the corresponding 4D theories, we find the IR dynamics of the 3D cases via compactification and a real mass deformation. We find that for the s-confining case with two antisymmetric fields, a second unlifted Coulomb branch direction indeed appears in the low- energy dynamics. We present several non-trivial consistency checks to establish the validity of these results. We also comment on the expected structure of general s-confining theories in 3D, which might involve several unlifted Coulomb branch directions. The original work presented in this thesis is mostly based on [1]. The paper will appear soon on JHEP. BIOGRAPHICAL SKETCH
Mario Martone is tremendously lucky of being born in the most beautiful and thrilling city in the world: Napoli, Italy. His loyalty and unconditional love for the city has been one of his most distinct traits in his adult life as an emigrant. Mario’s primary and secondary education took place in his beloved city where he attended the Scuola elementare Quarati to then move to the Scuola media Statale G. Lettieri. His love and interest for mathematics started at a very early age. At the first meeting with elementary school teachers, Mario’s math teacher told straight to his mother: signora, suo figlio ´eun genio! (madam, your son is a genius!). In his last year of elementary school Mario started play water-polo, another of his characteristic traits. He is now in his 21st consecutive year of playing the sport. After middle school Mario moved to the Liceo Classico A. Genovesi focusing his (non)-studies mostly on humanities, ancient Greek and Latin over anything else. During his high-school years he began his life-long commitment to revolutionary politics. During his junior year he was elected among the leaders of the school’s student movement and he and his amazing com- rades led multiple protests and take-overs of the school.
Although clearly a science oriented person, Mario didn’t develop his love for physics until the last year of high-school when he read Time Travel in Einstein’s Universe: The Physical Possi- bilities of Travel Through Time. He was completely fascinated by the deep philosophical impli- cations of modern physics and decided not to sell his soul by merely applying the beautiful laws of physics but try to contribute himself in the understanding of the fundaments of na- ture. He went on and pursued both a Laurea Triennale and a Laurea Specialistica (yes, Italy has a damn weird education system!) in physics. He was the first of his class to graduate obtaining a 110/110 cum laude (again Italy, why do you grade people out of 110?) with a thesis on the Geometrical Formulation of Quantum Mechanics. Fascinated by differential geometry but slightly thrown off from the lack of physical relevance of his undergraduate research, Mario moved to heart of the Empire to pursue a PhD at Syracuse University in Non-commutaive Geometry under the wonderful guidance of Prof. A. P. Balachandran.
iii In his first years of graduate career Mario spent extended amount time in Brazil visiting the International Center for Condensed Matter Physics (ICCMP), Universidade de Bras´ılia, then Spain visiting the Universidad Carlos III de Madrid and then India being a guest of the Institute of Math- ematical Sciences in Chennai. During these years Mario’s research focused on understanding the microscopic structure of space-time and its implications on space-time symmetries and quan- tum field theories. After his marvelous time in India, Mario enthusiastically accepted the kind offer from Prof. Yuval Grossman of visiting Cornell University for a semester and learn more about flavor physics. This semester long collaboration turned in pursuing his second PhD at
Cornell, this time in High Energy Physics. Mario completed his first PhD with honors in January 2012 with a thesis on Quantum Fields on Noncommutative Spacetimes. Since then he has finally settled his research interests on Beyond the Standard Model Physics and Formal Aspects of Quantum Field Theory. On June 9th, 2014 Mario’s beloved twin brother Alfonso had his first daughter Clara. Clara is undoubtedly the most amazing, intriguing and beautiful niece on planet Earth. Shortly after starting her journey outside of Sara’s womb, Clara committed to soon thoroughly study her new favorite book, incidentally bought to her by her uncle Mario: non-Euclidean geometry for Babies. Unfortunately at the moment, Clara is still a little too busy learning about the world to focus on math. Mario completed his second PhD in physics with the particle theory group at Cornell Uni- versity in August 2014 under the guidance of Prof. Csaba Csaki who, during the summer of 2014, was also initiated into the Napoli love. Mario is tremendously excited to continue his research in particle physics at the University of Cincinnati.
iv To the people. From Gaza to Ferguson, MO.
v ACKNOWLEDGEMENTS
First and foremost I want to thank my great friend and colleague Flip Tanedo without whom I would most likely be traveling around the world by now instead on working on physics. He helped me and encouraged me enormously during my academic-wise darkest times at Cornell. Our physics interests might be slightly divergent but I will always be there for you. Thanks man!
Secondly I want to deeply thank my adviser Prof. Csaba Csaki.´ I wish I had started col- laborating with him earlier in my career. Luckily enough I don’t see our collaboration ending anytime soon. Despite having very different personalities, our interactions have been incredi- bly insightful. You taught me a lot Csaba and I would love to keep learning from you! I want to thank Prof. Yuval Grossman without whom I would have never joined the Cornell Particle Theory group. He has also forced me once more to choose on which side of history to be on and what my life priorities really are. I want to thank UC Davis Prof. John Terning and UC Irvine Prof. Yuri Shirman for this last year of exciting collaborations...and the best is yet to come!
Then I want to thank the amazing people who accompanied my journey at Cornell. I was incredibly lucky to meet each one of them: Cody Long, down to Earth like few (especially in the Ivy world!) and always available for a chat/support/laugh/drinks, in three words an amazing friend, David Marsh, my only true physicist-comrade who stood always on my side, Paul McGuirk, an amazing person who lives life by example and for that deserves my deepest respect, I wish him the best of luck in his upcoming out-of-physics career, Riccardo Pavesi, one of the most talented physicists I have met with an incredibly loving and gentle heart, Dean J. Robinson, it is pretty amazing how far different we operate and yet how beautifully we worked together, Javi Serra, our collaboration is yet to bloom but I am sure it will soon, and finally my two beautiful office-mates Nic Rey-Le Lorier and Mathieu Cliche (Matt this acknowledgment dramatically underrepresents how much I am grateful to you for all the times you were there to help me). Thanks also to Brando Bellazzini, Josh Berger, Monika Blanke, Jack Collins, David
Curtin, Marco Farina, Naresh Kumar, Mike Saelim, Bibhushan Shakya, John Stout and Yuhsin
vi Tsai. And then there is my life outside of physics. First and foremost I want to thank my dearest
American comrades which luckily enough are too many to list. A special mention goes to the members of Student for Justice in Palestine who by being who they are have made the choice between my academic interests or fight for what I believe is just, a piece of cake. No matter how many threats I will receive, today, like that November 17th 2012, I will choose to stand and march by you! Also my Ithaca family: Ayslin, Clare, Gino, Kayla, James, Paula, Shawnae. Unfortunately the Shawn Greenwood Working Group is not going through its best days, but for sure we have done some incredible work together honoring the memory of Shawn Greenwood, a yet another young black man murdered by a Police officer. I love you all!
A massive thanks to all the members of the Cornell Water-polo team, ladies and boys alike. I will bring most of you in my heart for years to come. It has been an incredible journey to be involved with you for these last four years. Please continue with the champagne even without me! A special thanks to my closest US friends: Aisha, Anthony, Declan, Kevin, Julie, Max. Those countless drinks and amazing nights together contributed to make life in up-State NY so much better! Thanks to my girlfriend (WOW!) Ashley. The two of us have a hell lot to build together in Cincinnati.
And then there is my life outside of physics and outside of the US. Thank you to my family which represents still an incredibly solid pillar in my life despite thousands of miles among us. Thanks to Clara for being so incredibly cute. You contributed in giving me a purpose in life: make you the most spoiled niece ever. Finally I want to acknowledge the absolute centrality in my life of my napolitan brothers and comrades. Boccione, Carlo, Carolina, Claudia, Daniele, Gabriele, Germana, Julie, Luca,
Luca, Marcella, Pz8, Sciacallo, you are not only my closest friends and an infinite source of support and love. But you provide the reference against which I define myself and the choices I make in my life.
vii TABLE OF CONTENTS
Biographical Sketch...... iii Dedication...... v Acknowledgements...... vi Table of Contents...... viii List of Figures...... xi
1 Introduction 1
I General aspects of Coulomb Branches5
2 General aspects of N = 2 SUSY in Three Dimensions6 2.1 3D Spinors and the N = 2 SUSY Algebra...... 6 2.2 Chiral superfield...... 7 2.2.1 Real and Complex Masses...... 8 2.3 Vector superfield...... 9 2.3.1 Central Charge...... 10 2.4 Topological U(1)J symmetry...... 12 2.5 Linear superfield...... 13 2.6 Chern-Simons terms...... 15
3 Coulomb branch and the U(1) case 17 3.1 3D N = 2 U(1) gauge theory and moduli space of vacua...... 17 3.2 Coulomb branch and preliminary considerations...... 19 3.3 One-loop correction...... 20 3.4 Loop-induced global charges...... 23 3.5 U(1) with flavors...... 26 3.5.1 F = 1 ...... 27 3.5.2 F > 1 ...... 27 3.6 Mirror Symmetry...... 28
4 The non-abelian case 32 4.1 The non-abelian case...... 32 4.2 Coulomb branch pinch from matter...... 35 4.3 Instanton/Monopole background...... 37 4.3.1 BPS condition...... 38 4.3.2 Solution of the Bogomol’nyi equation...... 39 4.4 Non-perturbative corrections...... 40 4.5 Zero mode counting and general case...... 42 4.6 A taste of SU(N) dualities...... 46 4.6.1 F=N...... 47 4.6.2 F=N-1...... 47
viii 5 Monopole operators 48 5.1 BPS Vortices...... 49 5.1.1 Generic case...... 51 5.1.2 Fermionic zero-modes...... 52 5.2 A case of study: Mirror Symmetry...... 53 5.3 More general case...... 55 5.4 Non-abelian case...... 57
II The case with Antisymmetric matter 58
6 The Coulomb branch with antisymmetric matter 59 6.1 Coulomb branch coordinates...... 60 6.2 Dirac quantization condition of the monopole operators...... 62
7 Dimensional reduction of 4D dualities 66 7.1 General setting...... 66 7.2 S-confining case...... 68 7.2.1 Check 1...... 69 7.2.2 Check 2...... 70
8 Duals of 3D theories with antisymmetrics 72 8.1 3D duality for SU(4) with 2 A and 2 (Q + Q¯)...... 72 8.2 Consistency checks of the duality...... 76 8.2.1 Matching charges of the Coulomb branch operators...... 76 8.2.2 Matching the quantum constraints on a circle...... 77 8.3 Duality for SU(4) with A and 3(Q + Q¯) ...... 77 8.4 Low energy description on the Coulomb branch...... 78
9 Classification of 3D s-confining theories 80
10 Conclusions 83
A Spinors and notation 85
B Elements of Group theory 87 B.1 Definition and Cartan subalgebra...... 87 B.2 Roots, weights and lots of SU(2)s...... 88 B.2.1 weights...... 88 B.2.2 Adjoint representation and roots...... 88 B.2.3 Simple roots...... 90 B.2.4 SU(2)s...... 91 B.3 A closer look at SU(N) ...... 92 B.3.1 Fundamental...... 93 B.3.2 Antisymmetric...... 95
ix C Zero modes and Callias index theorem 97 C.1 Callias Index Theorem...... 97 C.2 Fundamental representation...... 97 C.3 Adjoint representation...... 98 C.4 Antisymmetric representation...... 98
Bibliography 100
x LIST OF FIGURES
3.1 In presence of matter, the Coulomb branch of the U(1) theory splits into two in- dependent branches and near the origin of the moduli space, where the Higgs branch intersects the Coulomb branches, the moduli space looks like the inter- section of three cones...... 22 3.2 Vacuum polarization diagram which generates one-loop Chern-Simons terms.. 24 3.3 Vacuum polarization diagram which generates one-loop mixed Chern-Simons terms. ai labels a vector associated to a weakly gauge U(1) theory...... 25
4.1 In presence of matter, the Coulomb branches pinch off where they meet Higgs branches. At these points, σ · να = 0 and there is massless matter...... 36 4.2 In the theory with fundamentals, the Coulomb branch splits at each point where σi = 0. The total Coulomb branch is composed of (N − 1) regions...... 43
7.1 If we reduce s-confining theories, there is no eηYe in the magnetic side...... 68
xi CHAPTER 1 INTRODUCTION
Quantum field theory is one of the richest subject in theoretical physics yet it still remains among the hardest to fully penetrate. Often in the last decades significant progresses in the understanding of gauge theory dynamics have come from looking at problems in simplified settings and only rarely by tackling straight the problem initially at hand. Supersymmetry and 3 dimensional physics, which represent the main focus of this thesis, have both being incredibly useful in providing simplified frames allowing considerable progress.
In the mid-1990s due to the seminal work of Seiberg and Intriligator on N = 1 theories [2–4], and Seiberg and Witten for N = 2 theories [5,6] the understanding of the dynamics of SUSY gauge theories was revolutionized. This brought to light a multitude of amazing properties which quantum field theory had: the interplay of holomorphy, global symmetries, instantons, anomalies, and monopoles can determine the IR behavior of a large class of theories. Depend- ing on the amount of matter, these theories manifest dynamical effects such as gaugino con- densation, instanton generated superpotentials, confinement with or without chiral symme- try breaking, IR-free composite gauge groups, interacting non-Abelian quantum fixed points, monopole condensation as a dual of confinement, etc.
On the other side, the study of lower-dimensional field theory represents a simpler labora- tory where to study and attempt to understand more complicated dynamics in 4D, for example, as a tool for understanding possible mechanisms for confinement [7,8]. More recently a new mechanism for confinement from bion condensation was proposed [9, 10] and further studied in [11–14]. Since one may generate 3D theories from the compactification of 4D theories, one can expect that some of their behavior is reflective of 4D properties. Furthermore, the study of lower-dimensional field theory has provided incredible insights in the understanding of dual- ities. For instance the equivalence of the massive Thirring and the sine-Gordon models [15, 16] has a very transparent meaning in that it can be shown that massive Thirring model arises re- writing the initial sine-Gordon model in terms of its topological solutions (kinks). This duality
1 is not simply conjectural as it can be shown that the operator which creates the sine-Gordon kink satisfy the same equation of motions as the massive Thirring models. It is believed that many of the higher dimensional conjectured dualities, arise in a similar fashion but thus far all attempts of formally writing down dualities for D > 2 have failed.
Extending the study to 3D N = 2 SUSY gauge theories many of the phenomena appearing in 4D have also been observed [17–35]. In particular the combination of SUSY with lower- dimensional field theory provides one of the most promising setting where an equivalence like the massive Thirren and sine-Gordon models could be provided. Mirror symmetry, which will be extensively reviewed in this thesis work, has in fact many of the features of lower di- mensional field theory [23, 56, 57]. The achievement of a formal duality map would provide a significant improvement in the understanding of gauge dualities.
Although we will touch on many of the described phenomena, our main focus in thesis will be slightly different. Over the past year, compelling derivations of many of the results obtained in [19–22,27,31,32] have emerged [36–38]. A careful sequence of compactification R4 → R3 × S 1 together with a real mass deformation yields purely three dimensional electric theories and allows one to determine the magnetic duals in a controlled way1 [36]. In this thesis we will present a first step to extend these results to theories with more general matter. In particular, we focus on the case with antisymmetric tensors. A careful counting of the fermionic zero modes in instanton backgrounds leads us to conclude that the dimension of the unlifted Coulomb branch can be larger once representations other than fundamentals are included. The reason for this is that the charges of the moduli parameterizing the Coulomb branch are modified in the presence of antisymmetric matter leading to fewer directions lifted by instanton effects. In the specific example of an SU(4) gauge group we identify two unlifted directions: the standard Y direction found in the presence of fundamental matter fields and a new Ye direction arising from the addi- tional splitting of the Coulomb branch due to massless components of the antisymmetric matter fields—this direction is related to the unlifted Coulomb branch modulus of SO(6) theories with vectors. Using this insight we apply the program of [36] to the simplest examples of 4D models
1See Chapter7 for a self-contained review of dimensional reduction of 4D dualities. Also see [39].
2 with antisymmetric matter and known 4D dynamics: the s-confining [40–42] SU(4) gauge the- ories with either two antisymmetrics and three flavors or one antisymmetric and four flavors.
We find that for the case of two antisymmetrics, the corresponding s-confining 3D theory with two flavors indeed requires the Ye operator in the description of the IR dynamics. Several con- sistency checks are presented, including the matching of quantum numbers, reproducing the quantum modified constraints expected for theories on a circle, and connecting it to theories with fewer flavors, establishing a consistent, intricate web of IR dynamics of several SU(4) the- ories with antisymmetric matter and flavors. The experience acquired with these theories leads us to speculate on the classification of the Coulomb branch structure for general s-confining theories in 3D.
The thesis splits naturally into two parts. In the first part we will mostly review known results. We will carefully and systematically present the material needed to tackle the analysis presented in the second part of the thesis. A review of N = 2 SUSY 3D theory is presented in Chapter2 where we outlined the differences between 3D N = 2 and 4D N = 1. Chapter 3 is dedicated to a thorough analysis of the dynamics of U(1) theories with a specific focus on the behavior of the Coulomb branch. The generalization to non-abelian theories (mostly
SU(N)) is performed in the subsequent chapter. There we will systematically explain the non- perturbative effects which drastically change the perturbative picture. Finally in Chapter5 we elaborate on how to interpret Coulomb branch operators across dualities. We will present a tentative a state-to-operator map between Coulomb branch operators and critical vortices.
In the second part of the thesis we will present the extension to the study of SU(N) theories with generic matter. Chapter6 contains a careful discussion of the structure of the Coulomb branch in the presence of antisymmetric matter. The counting of zero modes is performed and the general features outlined. In Chapter7 we will explain how to obtain 3D dualities from known 4D ones. This is mostly a review of [36]. We apply the program of [36] to the s-confining SU(4) theories with antisymmetric matter in Chapter8 to find the correct low-energy dynamics of those models, and perform various consistency checks. In Chapter9 we summarize our ex-
3 pectations for the general behavior of 3D s-confining theories, and finally conclude in Chapter 10. In the appendices we will review 3D spinors in AppendixA, provide elements of group theory in AppendixB. Zero mode counting and the Callias index theorem is presented in Ap- pendixC.
4 Part I
General aspects of Coulomb Branches
5 CHAPTER 2 GENERAL ASPECTS OF N = 2 SUSY IN THREE DIMENSIONS
In this first chapter we will introduce the formalism and the notation for the chapters to follow. In particular we will describe in details 3D N = 2 supersymmetry and its realization on super-space. For people already familiar with 4D SUSY, 3D N = 2 SUSY can be obtained by dimensional reduction of 4D N = 1 SUSY. Most of the material presented here follows the treatment in [19, 20, 43, 44]
2.1 3D Spinors and the N = 2 SUSY Algebra
A convenient representation of the Clifford algebra in 3D with metric ηi j = (−, +, +) is
i=0,1,2 γα,β = (iσ2, σ3, σ1). (2.1.1)
i j − i i jk Note that the generators of the Lorentz group, S = 2 γk, are pure imaginary and thus generate a real group. In fact in (2+1)-dimensions spinors come in representation of S pin(2, 1) ∼
SL(2, R), therefore the fundamental fermion representation is a 2-component Majorana fermion,
ψM. Under parity, this transforms as
P : ψM → ±γ1ψM. (2.1.2)
αβ As usual, spinor indices are contracted, raised and lowered with or αβ.
From the usual 4D N = 1 algebra, the 3D N = 2 SUSY algebra is
n o Qα, Qβ = [Pµ, Qα] = 0 n o ¯ i Qα, Qβ = 2γαβPi + 2iαβZ, (2.1.3) where the central charge Z is identified with momentum along xµ=2 in the 4D picture. The real- ization of (2.1.3) as differential operators acting on superspace follow from the 4D formalism:
∂ µ β ∂ β µ Qα = − iγ θ¯ ∂µ, Q¯α = − + iθ γ ∂µ (2.1.4) ∂θα αβ ∂θ¯α βα
6 Likewise the superspace derivatives, anti-commuting with (2.1.4), are ∂ ∂ D = + iγi θ¯β∂ , D¯ = − − iθβγi ∂ (2.1.5) α ∂θα αβ i α ∂θ¯α βα i From (2.1.5) one may read off the 3D N = 2 SUSY multiplets.
Finally we will define the following action of the differential operator (2.1.4) on a generic super field O h ¯i δζO ≡ ζQ O, δζ¯O ≡ ζ¯Q O (2.1.6)
Once an explicit form for the action (2.1.6) is provided, it possible to explicitly check the closure of the SUSY algebra and, as we will do below, compute explicitly the central charge value. In fact h i n o α β ¯ δζ, δζ¯ O = ζ ζ¯ Qα, Qβ O (2.1.7)
For completeness spinor notations is reviewed in AppendixA.
2.2 Chiral superfield
The first multiplet which we introduce is the chiral multiplet contained in the chiral super-field
Q which satisfies the constraint D¯ αQ = 0. Introducing the super-space coordinates (θ, θ,¯ y = x − iθγiθ¯) and using (2.1.5), we can derive an explicit expression for the chiral super-field
2 Q(y, θ, θ¯) = φQ(y) + θψ(y) + θ F(y) (2.2.1) where φQ is a complex scalar, ψ is a Weyl fermion which decomposes into two independent real Majorana fermions in 3D, and F is an auxiliary field.
From (2.1.4) and (2.1.6) we obtain the transformation properties of the chiral super-field
δζφ = ζψ, δζ¯φ = 0,,
i δζψ = ζF, δζ¯ψ = −iγ ζ∂¯ iφ, (2.2.2)
i δζ F = 0, δζ¯ F = −iζγ¯ ∂iψ .
7 The action for chiral superfields Q is Z Z 3 4 † 3 h 2 i S chiral = d x d θ K(Q, Q ) + d x d θ W(Q) + c.c. (2.2.3) for Kahler¨ potential K(Q, Q†) and superpotential W(Q). In particular for K(Q, Q†) = Q†Q the kinetic term is
2 † i 2 Lkin. = ∂iφQ + iψ γ ∂iψ + |F| , (2.2.4)
the subscript φQ indicates the lowest scalar component of the chiral superfield Q.
2.2.1 Real and Complex Masses
In 3D there are two different types of mass terms one may write for a chiral superfield Q. For a vector-like theory we can write down a holomorphic mass by adding a quadratic term to the super potential: W = m QQ¯ mC C (2.2.5) because mC is complex, this is known as a complex mass term and is the analog of the usual mass term in four dimensions. The mass term (2.2.5) is parity even.
Alternately, we can introduce a mass which is parity-odd. This is known as a real mass term and can be understood by modifying the Kahler¨ potential,
Z Z 2 ! 3 4 m θ2 † 3 m 2 αβ d xd θQe R Q ∼ d x R |φ | + im ψ¯ ψ . (2.2.6) 2 Q R α β
Again, observe that for the parity action (2.1.2), the complex mass preserves parity while the q m = m2 + m2 real mass breaks parity. The physical mass of the chiral superfield is R C.
A real mass can be induced by weakly gauging an exact global symmetry and fixing the weakly gauged vector superfield Vb into a SUSY-preserving background configuration (see be- low) m σ = i R , A = λ = λ¯ = D = 0. (2.2.7) b g i,b b b b
8 Because the global symmetries of dual theories must coincide, one may perform a real mass deformation on one theory and straightforwardly map it to the dual theory. This is used exten- sively in this paper and is explained thoroughly in Section8.
2.3 Vector superfield
The next multiplet is the vector multiplet which can be organized into a vector super-field satisfying V = V†. The action we will write down is invariant under super-gauge transformation
V → V + Λ + Λ†, for any chiral super-field Λ. In a particular choice of gauge, the Wess-Zumino gauge, one finds the following field content:
1 V = −iθθσ¯ − θγiθ¯A + iθ2θ¯λ¯ − iθ¯2θλ + θ2θ¯2D, (2.3.1) i 2
i where we have explicitly separated the 3D vector Ai from the gauge scalar σ ∼ A2 and the γ are defined in (2.1.1). Unlike the chiral superfield, the 3D N = 2 vector superfield differs from its 4D N = 1 counterpart in that it carries components which may acquire vacuum expectation values (VEVs) that form the Coulomb branch of the moduli space. We will discuss extensively Coulomb vacua throughout this thesis.
The superymmetry transformations on the vector super-field act as follow
i 1 δ A = − λ†γ ζ, δ σ = − λ†ζ ζ i 2 i ζ 2 i i jk δ λ = −γ F + D σ ζ, δ¯λ = 0. (2.3.2) ζ 2 i jk i ζ
A natural gauge invariant quantity which can be build out of the vector super-field V is
1 W = − D¯ 2e−V D eV (2.3.3) α 4 α
Acting with (2.1.5) it is straightforward to check that Wα is a chiral super field.
In terms of the the gauge-invariant combination (2.3.3), we can construct the supersymmet-
9 ric version of the Yang-Mills action
1 Z S = d3 x d2θ (Tr W Wα + c.c.) (2.3.4) YM g2 α Z ! 1 3 1 i j i 2 † i = d x Tr Fi jF + DiσD σ + D + λ γ Diλ . g2 4
We are now ready to write the supersymmetric action for a Yang-Mills theory with matter, that is coupling chiral super-fields to the vector super-fields. This is easily obtained modifying the Kahler potential K(Q, Q†) → Q†eV Q Z 3 4 † V S chYM = d xd θQ e Q (2.3.5) Z 3 2 † 2 † † i † † † † 2 = d x DiφQ + φQσ φQ + iφQDφQ + iψ γ Diψ − iψ σψ + iφQλ ψ‘ − iψ λφQ + |F| ,
where Di is the Dirac operator. From (2.3.5), we note that a non-vanishing VEV for the scalar component of the vector super-field, hσi , 0, induces a supersymmetric mass term for Q, specif- ically, a real-mass term as we already pointed out in (2.2.7).
After imposing back the Wess-Zumino gauge, the SUSY transformations (2.2.2) get modified to
δζφ = ζψ, δζ¯φ = 0,,
i δζψ = ζF, δζ¯ψ = −iγ ζ¯ Diφ − iσφ , (2.3.6)
i i δζ F = 0, δζ¯ F = −iζγ¯ − iγ Diψ + iσψ − iλφ .
2.3.1 Central Charge
Plugging (2.3.6) into (2.1.7) we can check that the SUSY algebra is closed up to a gauge transfor- mation. More interestingly if we consider a particular background configuration for the vector super field, namely ¯ σb = σeb, Ai,b = λb = λb = Db = 0 (2.3.7)
10 the action of (2.3.6) on the chiral super-field gives
h i α ¯β i δζ, δζ¯ Q = −ζ ζ 2iγαβ∂i + 2iαβσeb Q (2.3.8)
Thus σeb in (2.3.7) provides the value for the central charge. In other words, one can think of the central charge as lying in a background vector multiplet. Furthermore (2.3.7) is a SUSY preserv- ing background configuration as can be explicitly checked by looking at (2.3.2) and observing that SUSY transformations only depend on ∂iσ.
One can think of the background configuration (2.3.7) as gauging a particular flavor sym- metry and give it a very large kinetic term, so that it does not propagate and the field can be considered frozen in the given configuration. Because the kinetic term depends on the inverse of the gauge coupling 1/g2, this procedure is referred to as weakly gauging a flavor symmetry.
A weakly gauged flavor symmetry not only introduces a central charge but also a real mass for the chiral fields (2.2.7). The reverse can also be proven, that is any real mass in the theory can be interpreted as background configuration of a given weakly gauged flavor symmetry. This statement is of particular relevance for the work presented in this thesis. In fact we will only introduce real masses through weakly gauging global symmetries as this allows for a much simpler analysis when working with a dual pair. Knowing which background configura- tion provides a given real mass configuration in the electric theory, allows to straightforwardly guess the behavior of the magnetic side and the masses induced there .
q m ≥ Z m = m2 + m2 Massive representations satisfy , where in general C R. When the equality is satisfied the super-field is in a BPS configuration. It follows a very important result. In the absence of any mass term in the super-potential, that is if complex mass terms vanish, any chiral super-field charged under the weakly gauged flavor symmetry is in fact BPS.
Notice that because of the absence of anomalies, any global symmetry present in the theory can be weakly gauged.
11 2.4 Topological U(1)J symmetry
We have already mentioned that the existence of a scalar component σ in the vector super- multiplet, allows the possibility of a non-vanishing VEV for the vector super-field. At a more careful analysis, σ is not, however, the only scalar that can acquire a vacuum expectation value. In fact a 3D vector carries a single propagating degree of freedom and may be dualized into a scalar, γ.
In 3D, the Hodge dual of the field strength tensor, ?F, is a one-form which may locally be written with respect to a scalar γ,
[?F]i = dγ i . (2.4.1)
The field γ is known as the dual photon and encodes the degrees of freedom of Ai. Due to charge quantization, the dual photon γ is periodic and thus takes values on S 1. From this it follows that the topology of the 3D N = 2 Coulomb branch for a U(1) gauge theory is R × S 1. It is useful to combine the scalars into a complex modulus,
φ = σ + iγ. (2.4.2)
(2.4.2) is suggestive that the vector super-field can be dualized to a chiral super-field, with φ being its lowest complex scalar component. We will outline this construction in the next section.
The existence of a dual photon, also implies the presence of an extra U(1) global symmetry. Eq (2.4.1), together with Maxwell’s equation d ? F = 0, in fact implies that the ?F is divergence
jk free. We can thus define the following conserved current Ji ≡ i jkF . This current is associated with an extra U(1) global symmetry that, because of its nature, gets the name of topological U(1) or U(1)J . The U(1)J plays a very important role in studying the dynamics of the low energy physics in 3D.
12 2.5 Linear superfield
Lastly we can introduce the linear multiplet which can be arranged in the linear super-field Σ.
A linear super-field Σ satisfies D2Σ = D¯ 2Σ = 0 (2.5.1)
If written explicitly in terms of its components, the above condition can be understood as the
i supersymmetric generalization of current conservation ∂ Ji=0 [45]. Any divergence-free current
Ji is naturally included as one of the component of a linear super-field and, similarly, each linear super-field is associated to a symmetry of the theory whose conserved current is one of the component of the multiplet.
We have encountered a conserved current in the previous section where we showed that
i the dual field strength ?F = J is a divergence free one-form (2.4.1), ∂ Ji = 0, so that 3D gauge theories carry a global topological symmetry, U(1)J . We then expect that ?F should belong to a linear super-field Σ?F. This is in fact the case and the appropriate Σ?F can be directly constructed from the U(1) gauge theory vector super-field.
For each vector multiplet V, consider the new super-field Σ = DDV¯ . By construction Σ satisfies (2.5.1) and it is therefore a linear super-field. Explicitly, the linear super field associated to (2.3.1) is
i 1 i i 1 Σ ≡ − αβD¯ D V = σ + θλ¯ + θλ¯ + θγiθ¯J + iθθ¯D + θ¯2θγi∂ λ − θ2θγ¯ i∂ λ¯ + θ2θ¯2∂2σ. (2.5.2) 2 α β 2 i 2 i 2 i 4
The Σ introduced above contains ?F so it can be identified with Σ?F. Henceforth we will drop the extra label ?F and simply refer to the linear super-field containing dual field strength as Σ.
Using (2.5.2), we can write the YM action in terms of the linear super-field Σ
1 Z S = d3 xd4θ Σ2. (2.5.3) YM e2
Once expanded in terms of its component, (2.5.3) is completely equivalent to (2.3.4)
13 Linear super-field are extremely useful in 3D SUSY. We will see in the next section that they provide the natural supersymmetric generalization of Chern-Simons terms, but even more im- portantly the use of Σ instead of V allows a formulation of U(1) theories in terms of a Chiral super-field Φ. The complex scalar (2.4.2) may then be naturally understood as the lowest com- ponent of Φ which arises as the dual to the linear superfield (2.5.2). Let us see this.
The Lagrangian density which provides the YM action (2.5.3) can analogously be written as a function of a general real superfield Σ0 with the addition of a chiral superfield Φ acting as a
Lagrange multiplier to enforce the linear superfield conditions on Σ0, D2Σ0 = D¯ 2Σ0 = 0 ! Z Σ02 Σ0 L = d4θ − (Φ + Φ†) . (2.5.4) YM e2 2π
If the path integral over Σ0 is instead performed, one obtains a description of the physics in terms of the Chiral super-field Φ. Integrating out Σ0 leads to the condition
Σ0 Φ + Φ† = 4π (2.5.5) e2 from which one may write LYM as a function of the chiral super-field Φ. This is now a dual description of the vector superfield—encoded into a linear superfield—in terms of a chiral su- perfield whose lowest components are
† † Φ + Φ |θ=0 = 2Re[φ] ≡ 2ϕ Φ − Φ |θ=0 = 2iIm[φ] ≡ 2iγ. (2.5.6)
Using (2.5.5), the duality straightforwardly reproduces the claimed identification of (2.4.2) with the lowest component of Φ
2 jk ϕ ∼ 2πσ/e ∂iγ ∼ i jkF . (2.5.7)
In other words, the scalar component of the vector superfield is identified with the real part of the complex scalar in Φ, while its imaginary part can be identified with the dual photon.
14 2.6 Chern-Simons terms
In 3D, the Yang-Mills action is not the only gauge invariant combination of the gauge fields.
Chern-Simons terms can also be added to the action. While we do not discuss theories with non-vanishing Chern-Simons terms in any detail, they still play an important role in the theories discussed in this thesis as they are generated at one-loop. It is thus valuable to provide a brief discussion.
The supersymmetric generalization of Chern-Simons terms for generic non-abelian theories cannot be written down in super-space, yet it is possible to provide an explicit expression in terms of its component fields Z " # 3 i jk 2 † S = d x Tr A ∂ A + i A A A + 2Dσ − λ λ , (2.6.1) CS i j k 3 i j k
In the Abelian case, instead, supersymmetric CS terms can be written in a simple form in- volving both the vector and the linear multiplet: Z 3 4 S CS ≡ d x d θ ΣV. (2.6.2) such an expression can be generalized to mixed Chern-Simons terms Z Z mix 3 4 3 4 S CS ≡ d x d θ Σ1V2 = d x d θ Σ2V1 (2.6.3)
Mixed Chern-Simons terms play an important role in the discussion of the one-loop induced charges presented in the next chapter.
If any of the fields in (2.6.3) is taken to be in the background configuration described above, one obtains a Fayet-Iliopoulos term Z 3 4 S FI = ξ d xd θ V (2.6.4)
where ξ is the scalar component of the linear super-field Σb. This observation is very important. It shows that one possible way of obtaining a FI term is by weakly gauging the topological
U(1) instead of a flavor symmetry. Since the conserved U(1)J current is contained into the
15 linear super-field Σ, the U(1)J gauging introduces terms ∼ VJ Σ. VJ being the vector super-
field containing the photon for the now gauged U(1)J symmetry. If VJ is taken in a frozen background configuration, its scalar component provides a FI term. The relation between FI term and weakly gauging a U(1)J symmetry is thus analogous to the relation between central charges and weakly gauged flavor symmetries.
Before concluding this chapter is worth putting all the pieces together and explicitly write down the most general YM Lagrangian in 3D. For an abelian theory it has the form ! Z 1 k ξ L = d4θ − Σ2 − ΣV − V (2.6.5) e2 4π 2π
The first term being the tree level supersymmetric YM action, the second being a supersymmet- ric CS-term while the last a FI contribution. For non abelian theories ξ must be set to zero and the Chern-Simons term must be written explicitly in terms of the component fields as in (2.6.1).
In the next chapter we will provide details on how each one of terms above affect the physics.
16 CHAPTER 3 COULOMB BRANCH AND THE U(1) CASE
In this chapter we will start discussing in details the moduli space of N = 2 3D theories with particular attention to Coulomb branches as they will play a crucial role in the discussion outlined in subsequent chapters. In the U(1) case things are simpler. We thus start discussing the abelian case. Part of the material presented in this chapter can be found in [19, 20, 44]
It is important to notice that in 3D the gauge coupling is not dimensionless. U(1) theories, though simpler, are not IR free and they will present interesting dynamics which affects the topology of the moduli space. We will see that in the presence of matter quantum corrections drastically change the classical picture. But let’s start with describing the latter first.
3.1 3D N = 2 U(1) gauge theory and moduli space of vacua
As we saw in the previous chapter, the most generic Lagrangian for a U(1) gauge theory can be written as ! Z 1 k ξ L = d4θ − Σ2 − ΣV − V (3.1.1) e2 4π 2π where (3.1.1) is written in terms of both the vector (V) and linear super field (Σ) discussed extensively in the previous chapter.
The first contribution is the tree level gauge U(1) Lagrangian. The second term is the su- persymmetrised version of the Chern-Simons term. The last term is instead the more familiar Fayet-Iliopoulos term. In most of the examples that we will discuss, we will set both k = 0 and
ξ = 0 yet it is instructive to have a grasp on the effect of CS terms. In particular we will see below that despite being absent at tree level, CS terms could be generated radiatively at one loop.
Now consider adding matter fields Qi with U(1) charges ni and generic real masses mR,i.
17 Following [44], the semi-classical potential is 2 2 X X e 2 2 2 V = 2πni Qi − ζ − kσ + (m ,i + niσ) Qi (3.1.2) 32π2 R i i where σ is the lowest component of the linear super field (2.5.2). Here we set Wtree = 0. If a non- trivial super-potential is present, (3.1.2) is easily modified by adding the F-term contribution.
Consistency of the theory only allows certain value for the CS sector k 1 X k + n2 ∈ (3.1.3) 2 i Z i we will only work with theories where k is allowed to be set to zero. For more details, we refer to [20, 44].
Few observations are due. A non-vanishing value of the scalar component σ gives an effec- tive real mass to the matter fields Qi,
mi(σ) = mR,i + niσ. (3.1.4)
Thus for generic values of the real masses mR,i, Qi is massless only at σQi = −ni/mR,i. Away from
σQi , Qi is massive and can be integrated out generating one loop CS and FI terms (see below for more details). It follows that in (3.1.2) both k and ξ must be substituted by their effective values
1 X 2 k = k + n sign m + n σ (3.1.5) e f f 2 i R,i i i 1 X ξ = ξ + n m sign m + n σ (3.1.6) e f f 2 i i R,i i i
Semi-classical Supersymmetric vacua need to minimize (3.1.2), this implies X 2 2πni Qi = ξe f f + ke f f σ, &(mR,i + niσ)Qi = 0, for all i. (3.1.7) i
Solutions of (3.1.7) can be labeled as follows [44]:
1. Higgs vacua or Higgs branches, with hQii , 0 for some i. This implies that σ = σQi for
generic real masses mR,i. Below we will set mR,i = 0, then Higgs vacua are only allowed at the origin of the Coulomb branch (see below).
18 2. Coulomb vacua or Coulomb branches, solutions with σ , 0. In particular we will call Coulomb vacua only those solutions in which a continuous set of σ values is allowed. In particular
this implies hQii = 0 and ξe f f = ke f f = 0.
3. Topological vacua, cases in which hQii = 0 but only discrete values of σ are allowed. We will not discuss this case any further.
3.2 Coulomb branch and preliminary considerations
Let us recall the explicit expression for the 3D vector super-field
1 V = −iθθσ¯ − θγiθ¯A + iθ2θλ¯ − iθ¯2θλ + θ2θ¯2D, (3.2.1) i 2 where we have explicitly separated the 3D vector Ai from the gauge scalar σ ∼ A3. While in N = 1 4D case no vacuum expectation value is allowed for the vector super-field, N = 2 3D
SUSY admits Coulomb vacua. In fact if hQii = 0 and ke f f = ξe f f = 0, the σ component of the vector super-field (3.2.1) may acquire a vacuum expectation value (VEV).
Because of the existence of the dual photon γ (2.4.1) the topology of the Coulomb branch is classically R × S 1. The transformation properties under the previously introduced topological U(1) symmetry, suggest that quantum corrections modify this picture considerably. Let’s see this.
The U(1)J charge is defined as Z Z 2 0 2 F12 qJ = d x J = d x . (3.2.2) J 2π it then follows that in a U(1) gauge theory with F flavors, matter fields are invariant under the action of the topological U(1). The dual photon is instead charged under the U(1)J , and its action corresponds to a shift of γ and thus to going around the circle. A VEV of (2.4.2) breaks the U(1)J spontaneously with the dual photon arising as the Goldstone boson associated to the breaking.
19 We have already seen that classically the moduli space has a Coulomb branch with topology
R × S 1 and a Higgs branch that is parametrized by gauge neutral composite operators. The two branches intersect in one point, at σ = 0. At the intersection point the value of γ is un- constrained and we can move along the circle through the action of the U(1)J . On the other side the operators describing the Higgs branch are invariant under this action. It then follows that the intersection point of the Coulomb and the Higgs branch must be shrank to a point by quantum corrections, changing the topology of the Coulomb branch from the initial cylinder to two independent cones.
From its definition (3.2.2), qJ can be interpreted as magnetic charge, in 3D B ∼ F12. As we will describe in next chapters in more details, fields charged under the topological U(1) are in fact monopole operators and their action on the vacuum, with some non-trivial subtleties, create BPS vortices. For the moment we will not delve more into this connection and move to other properties of U(1) theories. We will come back often in subsequent sections to the connection between fields charged under the topological U(1) and monopole operators. We move now to compute explicitly the one-loop correction of the metric on the moduli space, providing a more formal derivation of the argument outlined above.
3.3 One-loop correction
In order to compute quantum correction on the moduli space metric, it is more useful to use a mixed variables basis: (σ, γ). The former being the lowest component of the linear super-field Σ and the latter the imaginary part of the lowest component of the chiral super-field Φ [19]. To account for quantum corrections we need to generalize slightly the duality described in Section 2.5.
The most general effective Lagrangian for a linear superfield is a real function of Σ Z 4 Leff = d θ f (Σ). (3.3.1)
20 and for a generic f (Σ),(2.5.5) generalizes to
∂ f (Σ0) Φ + Φ† = 2π (3.3.2) ∂Σ0 from which one may write now the Lagrangian in terms of the chiral super-field Φ Z 4 † Leff = d θ K(Φ + Φ ), (3.3.3) where K is the Legendre transform of f .
Including now one-loop contributions, the low-energy effective action reads ! Σ2 Σ f (Σ) = + Σ log (3.3.4) e2 e2
Plugging (3.3.4) back into (3.3.2), that is performing the Legendre transform, we thus obtain
Σ 1 1 Φ = + log(Σ/e2) + + iγ (3.3.5) 2e2 2 2
The metric on the moduli space is defined as
ds2 = K00(Φ + Φ¯ )dΦdΦ¯ . (3.3.6)
Thus we need to compute the second derivative of the Legendre transform in (3.3.3). This can be done noticing that the inverse of a Legendre transform is again a Legendre transform, or K0(Φ + Φ¯ ) = Σ01. Differetiating once more with respect to the chiral field Φ, we obtain
!−1 !−1 ∂(Φ + Φ¯ ) 1 1 K00(Φ + Φ¯ ) = = + (3.3.7) ∂Σ 2e2 Σ which can be plugged back into (3.3.6)
!−1 ! ! ! ! 1 1 1 1 1 1 1 1 ds2 = + + dσ + idγ + dσ − idγ . (3.3.8) 2e2 σ 2 2e2 σ 2 2e2 σ
(3.3.8) can be written in a more concise form as
! !−1 1 1 1 1 1 ds2 = + dσ2 + + dγ2 (3.3.9) 4 2e2 σ 2e2 σ
1In this paragraph we are using interchangeably the same letter for the fields and their VEVs, that is Φ ≡ hΦi.
21 Figure 3.1: In presence of matter, the Coulomb branch of the U(1) theory splits into two independent branches and near the origin of the moduli space, where the Higgs branch intersects the Coulomb branches, the moduli space looks like the intersection of three cones.
From (3.3.9), we can explicitly see the phenomenon outlined in the previous section. For large |σ| the metric (3.3.9) reduces to the canonical form. Yet σ, the scalar component of Σ, is no longer a good variable throughout the Coulomb branch. In fact once the one-loop correction is taken into account, the origin of the Coulomb branch, σ = 0, becomes a singularity. On the other side from (3.3.5) we can see that the chiral super field Φ becomes −∞ as we approach the singu- larity. It thus follows that the Coulomb branch of a U(1) theory splits into two holomorphically independent regions which are therefore parametrized by two chiral superfield V± ∼ exp ±Φ.
The expression of V± in terms of the chiral super-field Φ is only true in the semi-classical limit, that is for large VEVs and far away from the origin, this explains the sign ‘∼’. In other words
V± satisfy the constraint V+V− = 1 only semi-classically. Once quantum correction are taken into account, this constraint is modified and the two chiral fields become unconstrained V−V+ ∼ 0. Near the origin, where the Higgs branch and the Coulomb branch intersect, the moduli space looks like the intersection of three cones, see Fig. 3.1.
Both V+ and V− are charged under the topological U(1), while they are singlet under all the other global and gauge symmetries. From their semi-classical definition, it follows that they have qJ respectively +1 and −1. As we will describe in the next section, such operators can acquire charges under the other global U(1)’s through one-loop induced Chern-Simons terms.
22 3.4 Loop-induced global charges
One-loop induced U(1) charges are extensively used to understand the low-energy behavior of the 3D N = 2 theory, including in the non-abelian case. We will careful explain here how it works and the crucial role played by Chern-Simons terms in this construction.
The presence of a Chern-Simons term in the Lagrangian
µνρ µνρ LCS ∼ kTr Aµ∂νAρ = kTr AµFµν (3.4.1) modifies Gauss’ law (i.e. the A0 equation of motion) as follows 1 k − ∂ F = ρ − F (3.4.2) e2 i 0i matter 2π 12
δLmatter where ρmatter = is the matter contribution to the electric charge density, while a non- δA0 vanishing k value introduces a “magnetic” contribution.
The most practical way of getting rid of the gauge redundancy working within the canonical
2 scheme, is to impose the Coulomb gauge, i.e. A0 = 0. Thus A0 becomes a non-dynamical field and Gauss’ law has to be imposed as a constraint implying that any field with charge qJ acquires an electric charge
CS qelec = −kqJ. (3.4.3)
This effects it is also present for theories with no tree level CS term, that is if k = 0. This is because, as we have already mentioned a few times, Chern-Simons terms are generated radia- tivelly [20, 46, 47]. Let’s see now how this works in details.
If all real masses mR,i are set to zero, matter fields are only massless at σ = 0, while moving out in either direction of the Coulomb branch generates a real mass mi(σ) = niσ. It follows that when we compute the effective action for the light degrees of freedom for σ , 0, we need to integrate out these heavy fields. Then the vacuum diagram in Fig. 3.2 induces a term 1 Z d3q d3 p S [A] = n2 A (−p)A (−q)[−iΠlm(p)] (3.4.4) e f f 2 i (2π)3 (2π)3 l m
2 δL Recall that because of the antisymmetry of Fi j, A0 has no momentum conjugate, that is Π0 ≡ = 0. Thus δ∂0 A0 the Coulomb gauge allows to canonically quantize the other Aµ components, circumventing the problem.
23 A` Am
Figure 3.2: Vacuum polarization diagram which generates one-loop Chern-Simons terms.
lm where ni is the U(1) charge of the heavy fermion running in the loop and Π is defined as " # Z d3k −i −i Πlm(p) = −ie2 Tr γl γm (3.4.5) (2π)3 (p + k) − m(σ) k − m(σ)
The mass m(σ) really represents the real mass of the fermion running in the loop which is induced by moving out in the Coulomb branch, that is mi(σ) = niσ. In the formula above we suppressed the index to make the formula less cumbersome and avoid confusion.
For large m,
lm lmk Π → −i pk sign[m(σ)] + ... (3.4.6)
The dots indicate terms which need to be regularized. Plugging (3.4.6) back into (3.4.4), we see that each electrically charged fermion of the theory, generates an effective CS term at one loop
2 with coefficient keff,i = ni sign(niσ). Carefully integrating out all the massive fermions along the Coulomb branch we have 1 X k (σ) = k + n2sign(n σ) (3.4.7) eff 2 i i i where ni are the electric charges of matter fields. This result exactly reproduces the mR,i = 0 limit of (3.1.5).
From (3.4.7) and (3.4.3) we can read off the one-loop electric charge induced to the U(1)J charged field X CS 1 2 q = −k qJ = k + n sign(n σ) qJ . (3.4.8) elec e f f 2 i i i
In presence of non-vanishing real masses mR,i,(3.4.8) is easily generalized X X CS 1 2 1 2 q = k + n sign(m (σ)) qJ ≡ k + n sign(m + n σ) qJ . (3.4.9) elec 2 i i 2 i R,i i i i
24 a` Am
Figure 3.3: Vacuum polarization diagram which generates one-loop mixed Chern-Simons terms. ai labels a vector associated to a weakly gauge U(1) theory.
This formula gives the one-loop induced electric charge. In a similar way fields which have a non-vanishing qJ charge also acquire global U(1) charges at one-loop. One-loop global U(1) charges are generated by mixed Chern-Simons terms (2.6.3) which arise if we weakly gauge the particular U(1). Again let us see this explicitly.
Weakly gauging a particular U(1) global symmetry requires the introduction of an inter- action among the weakly gauged vector boson, that we will call ai, and the heavy fermions charged under such U(1). Thus when we integrate out heavy fields, loop diagrams involv- ing both Ai and ai (see Fig. 3.3) need to be also taken into account. Going through a similar calculation as before, they induce a one loop mixed Chern-Simons terms (2.6.3)
Lmixed ∼ mix i jk CS keff (σ)Tr aiF jk (3.4.10) where 1 X kmix(σ) = n n sign(n σ) (3.4.11) eff 2 iei i i eni representing the charge of the i-th matter field under the weakly gauged global U(1).
kmix U For eff , 0, the mixed CS terms will modify Gauss’ law for the weakly gauged (1) sym- metry (i.e. the a0 equation of motion), which in turns will induce a U(1) charge for fields which have a non-vaninishing qJ or F12 charge. Explicitly 1 X q = n n sign(n σ) qJ (3.4.12) e 2 iei i i
In summary, so far we have seen that U(1) theories with matter:
25 Have Higgs vacua parametrized by gauge invariant operators. • The Coulomb branch splits into two one-dimensional branches parametrized by two • quantum mechanically un-constrained operators which are given semi-classically by
V± = exp(±Φ).
In addition to the global symmetries of the initial theory there is an extra U(1), the U(1)J . • Matter fields are not charged under the topological U(1) while V± are.
Through one-loop generated CS terms, fields charged under the U(1)J , in particular V±, • acquire both electric (3.4.9) and global (3.4.12) U(1) charges.
In the next section we will see that the loop-induced charges play a crucial role to define the correct dynamics of the low-energy physics of U(1) theories with flavors.
3.5 U(1) with flavors
Consider a U(1) theory with F flavors. The quantum dynamics of such theories is constrained by the global quantum numbers of the fields
SU(F) SU(F) U(1)A U(1)J U(1)R
Q 1 1 0 0 (3.5.1) Q¯ 1 1 0 0
M ≡ QQ¯ 2 0 0 Furthermore the theory has two Coulomb branches described by the previously introduced operators V±. Both V+ and V− are charged under the topological U(1) thus they acquire charges under the global U(1) at one loop. From the table above and (3.4.12) we can read off the U(1) charges of the two Coulomb branch operators
U(1)A U(1)J U(1)R
V+ −F +1 F (3.5.2)
V− −F −1 F
26 The opposite charge for the U(1)R is due to the fact that the fermionic component in the chiral super-field has R-charge −1.
As already mentioned in previous sections, in 3D even the U(1) theory is not IR free, thus we expect the theory to flow to a non-trivial IR fixed point with a non-vanishing super-potential.
The super-potential has to respect all the global symmetries and have R-charge 2. This strongly constrains the possible dynamics.
3.5.1 F = 1
For F = 1 the Higgs branch and the two Coulomb branches are all complex manifolds of com- plex dimension 1. In this case the only Superpotential allowed by the symmetries in (3.5.1) and
(3.5.2) is
W = −MV+V− (3.5.3) thus SQED in 3D with one flavor flows to the XYZ model and the three cones which intersects near the origin of the two Coulomb branches are actually related by triality exchange symmetry.
3.5.2 F > 1
Now consider a U(1) gauge theory with a generic F. Again we can perform a similar analysis, using the symmetries in (3.5.1) and (3.5.2) to infer the low energy behavior. This gives the super-potential
1/F W = −F(V+V− det M) (3.5.4) this super-potential provides the correct description of the low-energy theory away from the origin but it provides a singular scalar potential at the origin. This suggests that we need ad- ditional degrees of freedom to be able to describe the theory in a generic point of the moduli space.
27 This description will be provided in the next section using results from mirror symmetry. A careful analysis of U(1) theories provides a good background to understand the behavior of the Coulomb branch in non-abelian theory with flavors which will be described in the next chapter. In the second part of this thesis we will further generalize the analysis to theories with antisymmetric tensors. Furthermore, mirror symmetry also sheds light in the connection between monopole operators and Coulomb branches. This will be addressed in Chapter5.
3.6 Mirror Symmetry
3D U(1) theories with F fundamentals in N = 4 have been studied and understood by a series of beautiful papers [18, 23, 48–50]. They are a subset of a bigger list of theories which show a very peculiar kind of duality named “Mirror symmetry”. N = 4 theories in 3D have a global
SO(4) SU(2)L×SU(2)R, with SU(2)R the R symmetry of N = 1 in 6D or N = 2 in 4D, and SU(2)L associated with rotations in the three directions reduced in going from 6 to 3 dimensions. Mirror symmetry exchanges [18]:
SU(2)L and SU(2)R. • The Coulomb and the Higgs branch. • Mass and Fayet-Iliopoulos terms. •
By promoting some of the 3D N = 4 coupling constants to background super-fields and analyzing their transformation properties under the global SO(4), it is possible to prove a few non-renormalization theorems. Namely that [18]
Higgs branches are not renormalized by quantum effects while Coulomb branches are. • Mass terms only affect the metric on the Coulomb branch. • Fayet-Ilipoulos terms only affect the metric on the Higgs branch instead. •
28 Mirror symmetries are very interesting dualities. In fact from the properties above and the inter-change of the Higgs/Coulomb branch, it follows that the quantum modifications of the moduli space in one side of the duality should be already present at the classical level in the other side. We will see examples of this phenomenon shortly, for instance the extensively discussed V± splitting of the Coulomb branch of U(1) theories it is evident classically looking at the Higgs branch of the Mirror dual. In the following chapters we will also see that Mirror symmetry can be understood as particle-vortex duality and we will explicitly build the map. This will provide a case study to see in action how Coulomb branch operators could show up as particle-like operators in the dual side. Therefore it is important to discuss these dualities in details in order to lay the ground for the subsequent discussion on monopole operators and to understand their interpretation across the duality. Also by breaking N = 4 to N = 2, we can use mirror symmetry to provide the low energy description of U(1) theories with more than a single flavor.
Flowing from N = 4 to N = 2 in 3D is in many ways analogous to the flow from N = 2 to N = 1 in 4D. This can be achieved noticing that the N = 4 vector multiplet contains a N = 2 vector and chiral multiplet while N = 4 hyper-multiplets contain a chiral and an anti-chiral
N = 2 multiplets. Giving a mass to Ψ, the chiral component of the N = 4 vector multiplet
i) Breaks N = 4 to N = 2.
ii) The chiral component of the N = 4 vector multiplet decouples leaving only the N = 2 vector multiplet in the low-energy theory.
iii) Each hypermultiplets now should be read as a flavor of N = 2.
Thus a mass for Ψ is enough to flow exactly to the kind of N = 2 theories which we are interested in. U(1) theories with F flavors in N = 4 are known to have a “mirror dual” in terms