Collective T-duality transformations and non-geometric spaces

Similar documents
Towards new non-geometric backgrounds

Relating DFT to N=2 gauged supergravity

Non-Associative Flux Algebra in String and M-theory from Octonions

T-duality : a basic introduction

Double Field Theory and Stringy Geometry

Flux Compactification of Type IIB Supergravity

Branes in Flux Backgrounds and Dualities

Non-associative Deformations of Geometry in Double Field Theory

Instantons in string theory via F-theory

Double Field Theory at SL(2) angles

U-dual branes and non-geometric string backgrounds

Extended Space for. Falk Hassler. bases on. arxiv: and in collaboration with. Pascal du Bosque and Dieter Lüst

Cosmological solutions of Double field theory

F-theory effective physics via M-theory. Thomas W. Grimm!! Max Planck Institute for Physics (Werner-Heisenberg-Institut)! Munich

The Geometry behind Nongeometric Fluxes

C-spaces, Gerbes and Patching DFT

Non-Commutative/Non-Associative Geometry and Non-geometric String Backgrounds

Open String Wavefunctions in Flux Compactifications. Fernando Marchesano

Coset CFTs, high spin sectors and non-abelian T-duality

Non-commutative & non-associative closed string geometry from flux compactifications

Flux Compactifications and Matrix Models for Superstrings

String Geometry Beyond the Torus

Gravity theory on Poisson manifold with R-flux

Three-Charge Black Holes and ¼ BPS States in Little String Theory I

Heterotic Flux Compactifications

Techniques for exact calculations in 4D SUSY gauge theories

Branes, Wrapping Rules and Mixed-symmetry Potentials

Membrane σ-models and quantization of non-geometric flux backgrounds

Think Globally, Act Locally

2. Lie groups as manifolds. SU(2) and the three-sphere. * version 1.4 *

Heterotic Torsional Backgrounds, from Supergravity to CFT

Non-Associative Geometry and Double Field Theory

R-flux string sigma model and algebroid duality on Lie 3-algebroids

T-duality & noncommutative geometry

SUPERSTRING REALIZATIONS OF SUPERGRAVITY IN TEN AND LOWER DIMENSIONS. John H. Schwarz. Dedicated to the memory of Joël Scherk

D-branes as a single object. SIS Dubna, Edvard Musaev

A Geometry for Non-Geometric String Backgrounds

Generalized N = 1 orientifold compactifications

MIFPA PiTP Lectures. Katrin Becker 1. Department of Physics, Texas A&M University, College Station, TX 77843, USA. 1

8.821 String Theory Fall 2008

8.821 F2008 Lecture 12: Boundary of AdS; Poincaré patch; wave equation in AdS

q form field and Hodge duality on brane

Higher Structures in Non-Geometric M-Theory

Quantum Nambu Geometry in String Theory

Fabio Riccioni. 17th July 2018 New Frontiers in String Theory 2018 Yukawa Institute for Theoretical Physics, Kyoto

8.821 String Theory Fall 2008

Dualities in field theories and the role of K-theory

Double Field Theory Double Fun?

What is F-theory? David R. Morrison. University of California, Santa Barbara

Cold atoms and AdS/CFT

Three-Charge Black Holes and ¼ BPS States in Little String Theory

Heterotic Geometry and Fluxes

Glimpses of Double Field Theory Geometry

The Conformal Algebra

Half BPS solutions in type IIB and M-theory

A Brief Introduction to AdS/CFT Correspondence

Little strings and T-duality

BMS current algebra and central extension

Non-relativistic holography

Supergeometry and unified picture of fluxes

Dimensional reduction

On Flux Quantization in F-Theory

M-Theory and Matrix Models

Microscopic Formulation of Puff Field Theory. Akikazu Hashimoto. University of Wisconsin Madison, February 2007

Contents. Preface to the second edition. Preface to the first edition. Part I Introduction to gravity and supergravity 1

Think Globally, Act Locally

SCFTs, Compact CY 3-folds, and Topological Strings

Non-relativistic AdS/CFT

Symmetry and Geometry in String Theory

Black hole near-horizon geometries

Non-Geometric Calabi- Yau Backgrounds

David R. Morrison. String Phenomenology 2008 University of Pennsylvania 31 May 2008

Three-dimensional gravity. Max Bañados Pontificia Universidad Católica de Chile

New Frontiers in String Theory 2018 Yukawa Institute for Theoretical Physics (YITP) Kyoto - July, 12th 2018

Knot Homology from Refined Chern-Simons Theory

Geometry and Physics. Amer Iqbal. March 4, 2010

New Phenomena in 2d String Theory

Super p-brane Theory. Super Homotopy Theory

Elliptic Genera of non-compact CFTs

Dynamics of branes in DFT

The correspondence between free fermionic models and orbifolds

Moduli of heterotic G2 compactifications

BPS non-local operators in AdS/CFT correspondence. Satoshi Yamaguchi (Seoul National University) E. Koh, SY, arxiv: to appear in JHEP

4. Killing form, root space inner product, and commutation relations * version 1.5 *

Brief course of lectures at 18th APCTP Winter School on Fundamental Physics

Lecture 8: 1-loop closed string vacuum amplitude

arxiv:hep-th/ v3 21 Jul 1997

Entropy of asymptotically flat black holes in gauged supergravit

Fourth Aegean Summer School: Black Holes Mytilene, Island of Lesvos September 18, 2007

ELEVEN DIMENSIONS FROM THE MASSIVE D-2-BRANE

A Landscape of Field Theories

Candidates for Inflation in Type IIB/F-theory Flux Compactifications

Anomalous discrete symmetries in D-brane models

Warped Models in String Theory

Spectrum of Holographic Wilson Loops

MAC2313 Final A. (5 pts) 1. How many of the following are necessarily true? i. The vector field F = 2x + 3y, 3x 5y is conservative.

arxiv:hep-th/ v2 14 Oct 1997

M-theory and extended geometry

LECTURE 5: SURFACES IN PROJECTIVE SPACE. 1. Projective space

Superstring in the plane-wave background with RR-flux as a conformal field theory

Transcription:

Collective T-duality transformations and non-geometric spaces Erik Plauschinn LMU Munich ESI Vienna 09.12.2015

based on... This talk is based on :: T-duality revisited On T-duality transformations for the three-sphere and some work in progress [arxiv:1310.4194] [arxiv:1408.1715] [arxiv:15xx.xxxxx]

motivation :: dualities I Duality :: two different theories describe the same physics.

motivation :: dualities I Duality :: two different theories describe the same physics. They play an important role in understanding string theory. M-theory Type IIA T Type IIB S Het E 8 E 8 T Het SO(32) S Type I

motivation :: dualities II New solutions can be found by applying duality transformations. known solution duality new solution

motivation :: dualities II New solutions can be found by applying duality transformations. known solution duality new solution In particular, consider string theory :: geometric space T-duality non-geometric space

motivation :: dualities II New solutions can be found by applying duality transformations. known solution duality new solution In particular, consider string theory :: geometric space T-duality non-geometric space

motivation :: goal Main goal :: understand better non-geometric spaces

motivation :: goal study (collective) T-duality transformations Main goal :: understand better non-geometric spaces

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples 5. summary

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples 5. summary

outline 1. motivation 2. non-geometric spaces a) t-duality b) non-geometry c) questions 3. collective t-duality 4. examples 5. summary

outline 1. motivation 2. non-geometric spaces a) t-duality b) non-geometry c) questions 3. collective t-duality 4. examples 5. summary

t-duality :: circle Consider the torus partition function of the bosonic string (with = 1 + i 2 ) Z( 1, 2 )=Tr H e 2 2H e +2 1P. For a compactification on a circle with radius R, it reads Z(, ) = 1 p 2 23 1 24 ( ) 2 X m,n2z exp + i apple m 2 0 R + Rn 0 2! exp apple i m 2 0 R Rn 0 2!. This partition function is invariant under the replacement R! 0 R.

t-duality :: circle Consider the torus partition function of the bosonic string (with = 1 + i 2 ) Z( 1, 2 )=Tr H e 2 2H e +2 1P. For a compactification on a circle with radius R, it reads Z(, ) = 1 p 2 23 1 24 ( ) 2 X m,n2z exp + i apple m 2 0 R + Rn 0 2! exp apple i m 2 0 R Rn 0 2!. This partition function is invariant under the replacement R! 0 R.

t-duality :: torus T-duality can be extended from the circle to toroidal backgrounds. R (G,B) For constant metric G and Kalb-Ramond field B, the partition function is invariant under O(D,D) transformations of H = G BG 1 B +BG 1 G 1 B G 1!.

t-duality :: curved background For general CFT backgrounds, with non-constant metric and B-field, one considers Z 1 h S = 4 0 G ij dx i ^?dx j + ib ij dx i ^ dx j + 0 R?1i. Gauging an isometry and integrating out the gauge field, gives the Buscher rules G = 1 G, G a = B a G, B a = G a G, G ab = G ab G a G b B a B b G, B ab = B ab G a B b B a G b G. The dual theory is conformal, if the dilaton is shifted as = + 1 2 log r det G det G. Buscher - 1985, 1987, 1988

t-duality :: summary T-duality :: circle compactification R! 1/R, toroidal background curved background O(D,D) transformation, Buscher rules.

outline 1. motivation 2. non-geometric spaces a) t-duality b) non-geometry c) questions 3. collective t-duality 4. examples 5. summary

non-geometry Non-geometry :: apply successive T-duality to a three-torus with H-flux.

non-geometry H abc T c! fab c T b! Qa bc T a! R abc

non-geometry :: h-flux H abc T c! fab c T b! Qa bc T a! R abc Consider string theory compactified on a three-torus with H-flux: The geometry is determined by ds 2 = dx 2 + dy 2 + dz 2, B yz = Nx, x x +1, y y +1, z z +1. The H-flux reads H xyz = N.

non-geometry :: h-flux H abc T c! fab c T b! Qa bc T a! R abc Consider string theory compactified on a three-torus with H-flux: The geometry is determined by ds 2 = dx 2 + dy 2 + dz 2, B yz = Nx, x x +1, y y +1, z z +1. The H-flux reads gauge tr. H xyz = N. (G, B)(x + 1) (G, B)(x) x

non-geometry :: h-flux H abc T c! fab c T b! Qa bc T a! R abc Consider string theory compactified on a three-torus with H-flux: The geometry is determined by ds 2 = dx 2 + dy 2 + dz 2, B yz = Nx, x x +1, y y +1, z z +1. The H-flux reads H xyz = N.

non-geometry :: f-flux H abc T c! fab c T b! Qa bc T a! R abc After a T-duality in the z-direction, one arrives at a twisted torus: The geometry is determined by ds 2 = dx 2 + dy 2 +(dz + Nxdy) 2, B =0, (x, z) (x +1,z Ny), y y +1, z z +1. The geometric flux arises via e x = dx, e y = dy, e z = dz + Nxdy,! z xy = N/2, [e x,e y ]= N e z. Scherk, Schwarz - 1979 Dasgupta, Rajesh, Sethi - 1999 Kachru, Schulz, Tripathy, Trivedi - 2002

non-geometry :: f-flux H abc T c! fab c T b! Qa bc T a! R abc After a T-duality in the z-direction, one arrives at a twisted torus: The geometry is determined by ds 2 = dx 2 + dy 2 +(dz + Nxdy) 2, B =0, (x, z) (x +1,z Ny), y y +1, z z +1. The geometric flux arises via e x = dx, e y = dy, e z = dz + Nxdy, (G, B)(x + 1) SL(2, Z) (G, B)(x)! z xy = N/2, [e x,e y ]= N e z. x Scherk, Schwarz - 1979 Dasgupta, Rajesh, Sethi - 1999 Kachru, Schulz, Tripathy, Trivedi - 2002

non-geometry :: f-flux H abc T c! fab c T b! Qa bc T a! R abc After a T-duality in the z-direction, one arrives at a twisted torus: The geometry is determined by ds 2 = dx 2 + dy 2 +(dz + Nxdy) 2, B =0, (x, z) (x +1,z Ny), y y +1, z z +1. The geometric flux arises via e x = dx, e y = dy, e z = dz + Nxdy,! z xy = N/2, [e x,e y ]= N e z. Scherk, Schwarz - 1979 Dasgupta, Rajesh, Sethi - 1999 Kachru, Schulz, Tripathy, Trivedi - 2002

non-geometry :: q-flux H abc T c! fab c T b! Qa bc T a! R abc After a second T-duality in the x-direction, one arrives at a T-fold: The geometry is determined by ds 2 = dy 2 1 + 1+(Ny) 2 (dx2 + dz 2 ), Ny B xz = 1+(Ny) 2, y y +1, z z +1. The non-geometric flux reads Q y xz = N. This space is locally geometry, but globally non-geometric. Hellermann, McGreevy, Williams - 2002 Dabholkar, Hull - 2002 Hull - 2004

non-geometry :: q-flux H abc T c! fab c T b! Qa bc T a! R abc After a second T-duality in the x-direction, one arrives at a T-fold: The geometry is determined by ds 2 = dy 2 1 + 1+(Ny) 2 (dx2 + dz 2 ), Ny B xz = 1+(Ny) 2, y y +1, z z +1. The non-geometric T-duality flux reads (G, B)(y + 1) Q y xz = N. (G, B)(y) This space is locally geometry, but globally non-geometric. y Hellermann, McGreevy, Williams - 2002 Dabholkar, Hull - 2002 Hull - 2004

non-geometry :: q-flux H abc T c! fab c T b! Qa bc T a! R abc After a second T-duality in the x-direction, one arrives at a T-fold: The geometry is determined by ds 2 = dy 2 1 + 1+(Ny) 2 (dx2 + dz 2 ), Ny B xz = 1+(Ny) 2, y y +1, z z +1. The non-geometric flux reads Q y xz = N. This space is locally geometry, but globally non-geometric. Hellermann, McGreevy, Williams - 2002 Dabholkar, Hull - 2002 Hull - 2004

non-geometry :: r-flux H abc T c! fab c T b! Qa bc T a! R abc After formally applying a third T-duality, one obtains an R-flux background: The geometry is not even locally defined. The non-geometric R-flux is obtained by raising the index of the Q-flux Q y xz! R xyz = N. This background gives rise to a non-associative structure. Bouwknegt, Hannabuss, Mathai - 2004 Shelton, Taylor, Wecht - 2005 Ellwood, Hashimoto - 2006 Blumenhagen, EP & Lüst - 2010

non-geometry :: topological t-duality T-duality (toroidal) has also been studied in a more mathematical approach. three-torus with H-flux 1 T-duality twisted torus = S 1 E T 2 2 T-dualities non-commutative torus 3 T-dualities non-associative torus Bouwknegt, Evslin, Mathai - 2003 Mathai, Rosenberg - 2004 Bouwknegt, Hannabuss, Mathai - 2005

non-geometry :: topological t-duality T-duality (toroidal) has also been studied in a more mathematical approach. three-torus with H-flux 1 T-duality twisted torus = S 1 E T 2 2 T-dualities non-commutative torus 3 T-dualities non-associative torus Bouwknegt, Evslin, Mathai - 2003 Mathai, Rosenberg - 2004 Bouwknegt, Hannabuss, Mathai - 2005 what is the precise relation to string theory?

non-geometry :: summary Summary :: non-geometry arises through a chain of T-duality transformations. H abc T c! fab c T b! Qa bc T a! R abc flux background "twisted torus" T-fold non-associative

outline 1. motivation 2. non-geometric spaces a) t-duality b) non-geometry c) questions 3. collective t-duality 4. examples 5. summary

open questions Questions :: 1. Why are three T-dualities (three-torus) not allowed? 2. Can other examples be constructed? 3. "What is the origin of non-geometry?"

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples 5. summary

collective t-duality :: main idea Main idea :: 1. Identify a global symmetry of the theory. 2. Gauge this symmetry. 3. Integrate-out the gauge field. gauged theory gauging AS =0 original theory dual theory Buscher - 1985, 1987, 1988

collective t-duality :: main idea Main idea :: 1. Identify a global symmetry of the theory. 2. Gauge this symmetry. 3. Integrate-out the gauge field. gauged theory da =0 gauging AS =0 original theory dual theory Buscher - 1985, 1987, 1988

collective t-duality :: sigma-model Consider the sigma-model action for the NS-NS sector of the closed string S = 1 4 0 Z @ h i G ij dx i ^?dx j + 0 R?1 i 2 0 Z 1 3! H ijk dx i ^ dx j ^ dx k. This action is invariant under global transformations X i = k i (X) if L k G =0, k H = dv, L k =0. In general, the isometry algebra is non-abelian :: [k,k ] L = f k. Hull, Spence - 1989 & 1991

collective t-duality :: gauged action Following Buscher s procedure, one first gauges isometries of the action (ignore dilaton) bs = 1 2 0 i 2 0 i 2 0 Z Z Z @ @ h 1 2 G ij(dx i + k i A ) ^?(dx j + k j A ) 1 3! H ijk dx i ^ dx j ^ dx k (v + d ) ^ A + 1 2 k [ v ] + f A ^ A i. Hull, Spence - 1989 & 1991 Alvarez, Alvarez-Gaume, Barbon, Lozano - 1994 The local symmetry transformations and constraints take the form ˆ X i = k i, ˆ A = d A f, ˆ = k( v ) f, L k[ v ] = f v, 3 k[ f ] v = k k k H. EP - 2014

collective t-duality :: original model The original model is recovered via the equations of motion for 0=dA 1 2 f A ^ A. The gauged action can then be rewritten in terms of bs = 1 4 0 i 2 0 Z Z @ DX i = dx i + k A i as h i G ij DX i ^?DX j + 0 R?1 1 3! H ijk DX i ^ DX j ^ DX k. Ignoring technical details, one replaces DX i! dy i and obtains the ungauged action.

collective t-duality :: dual model I The dual model is obtained via the equations of motion for A A = G DG 1 D 1 1 + i? DG 1 k + i?, where G = k i G ij k j, = d + v, D = k[ v ] + f, k = k i G ij dx j. The action of the dual sigma-model is found by integrating-out Z Z 1 Š = hǧ + 0 i 4 0 R?1i 2 0 Ȟ, @ A and reads where, with M = G DG 1 D Ǧ = G + invertible, T k M 1 M 1 DG 1 +M 1 DG 1 +M 1 k ^?, Ȟ = H + 1 2 d " k T +M 1 DG 1 +M 1 M 1 M 1 DG 1 ^ # k.

collective t-duality :: dual model II The coordinates X i and fields parametrize an enlarged target-space. The enlarged metric Ǧ and field strength Ȟ have null-eigenvectors ň Ǧ =0, ň Ȟ =0, ň = k + D @. The dual metric and field strength are obtained via a change of coordinates T I A = k 0, Ǧ D 1 AB =(T T Ǧ T ) AB = 0 0, 0 G Ȟ ABC = ȞIJK T I A T J B T K C, Ȟ ibc =0.

collective t-duality :: summary The T-duality transformation rules are obtained via Buscher s procedure of 1. gauging isometries in the sigma-model action, 2. integrating-out the gauge field, 3. performing a change of coordinates. The following restrictions arise in this procedure :: from gauge invariance (recall that k H = dv ) L k[ v ] = f v, k[ f ] v = 1 3 k k k H. from solving equations of motion (recall that G = k G i ij k j ) det G 6= 0, det G DG 1 D 6= 0. The change of coordinates is performed using null-eigenvectors.

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples 5. summary

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples a) three-torus b) three-sphere 5. summary

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples a) three-torus b) three-sphere 5. summary

torus :: setting Consider a three-torus with H-flux specified as follows ds 2 = R 2 1 dx 1 2 + R 2 2 dx 2 2 + R 2 3 dx 3 2, X i ' X i + `s, H = hdx 1 ^ dx 2 ^ dx 3, h 2 ` 1 s Z. The Killing vectors of interest are abelian and can be chosen as k 1 = @ 1, k 2 = @ 2, k 3 = @ 3. The one-forms v are defined via k H = dv.

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. T k M Ǧ = G + 1 M 1 DG 1 k +M 1 DG 1 +M 1 ^? The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as = d + v

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. T k M Ǧ = G + 1 M 1 DG 1 k +M 1 DG 1 +M 1 ^? The constraints for gauging the sigma-model are trivially satisfied. = G + R 2 1 dx 1 T 1 R 2 1 The geometry of the dual background is determined from the quantities 0 0 + 1 R 2 1! R 2 ^? 1 dx 1 G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as = d + v

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. T k M Ǧ = G + 1 M 1 DG 1 k +M 1 DG 1 +M 1 ^? The constraints for gauging the sigma-model are trivially satisfied. = G + R 2 1 dx 1 The geometry of the dual background is determined from the quantities = G R G = R1 2 1 2 dx 1 ^?dx 1 + 1, = d + v, R1 2 D =0, k = R 2 1 dx 1, The metric and field strength are obtained as T 1 R 2 1 0 0 + 1 R 2 1! ^? R 2 ^? 1 dx 1 M = G = R 2 1. = d + v

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. T k M Ǧ = G + 1 M 1 DG 1 k +M 1 DG 1 +M 1 ^? The constraints for gauging the sigma-model are trivially satisfied. = G + R 2 1 dx 1 The geometry of the dual background is determined from the quantities = G R G = R 2 1 2 dx 1 ^?dx 1 + 1 1, = d + v, R 2 ^? 1 M = G = R1 2. D =0, = 1 k = 1 2 dx 1, R1 2 ^? + R2 2 dx 2 ^?dx 2 + R3 2 dx 3 ^?dx 3 The metric and field strength are obtained as T 1 R 2 1 0 0 + 1 R 2 1! R 2 ^? 1 dx 1 = d + v

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. Ȟ = H + 1 2 d " k T +M 1 DG 1 +M 1 The constraints for gauging the sigma-model are trivially satisfied. M 1 M 1 DG 1 ^ # k The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. Ȟ = H + 1 2 d " k T +M 1 DG 1 +M 1 The constraints for gauging the sigma-model are trivially satisfied. M 1 M 1 DG 1 ^ # k = H + 1 2 d " R 2 1 dx 1 T 0 + 1 R1 2 1 0 The geometry of the dual background is determined from the quantities R 2 1! ^ R 2 1 dx 1 # G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. Ȟ = H + 1 2 d " k T +M 1 DG 1 +M 1 The constraints for gauging the sigma-model are trivially satisfied. M 1 M 1 DG 1 ^ # k = H + 1 2 d " R 2 1 dx 1 The geometry of the dual background is determined from the quantities h i G = R= 1 2, H + d dx 1 ^= d + v, D =0, k = R 2 1 dx 1, The metric and field strength are obtained as T 0 + 1 R1 2 1 0 R 2 1! ^ R 2 1 dx 1 # M = G = R 2 1.

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. Ȟ = H + 1 2 d " k T +M 1 DG 1 +M 1 The constraints for gauging the sigma-model are trivially satisfied. M 1 M 1 DG 1 ^ # k = H + 1 2 d " R 2 1 dx 1 The geometry of the dual background is determined from the quantities h i G = R= 1 2, H + d dx 1 ^= d + v, D =0, =0 k = R1 2 dx 1, The metric and field strength are obtained as T 0 + 1 R1 2 1 0 R 2 1! ^ R 2 1 dx 1 # M = G = R 2 1. d = d(d + v) =hdx 2 ^ dx 3

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as

torus :: one t-duality I Consider one T-duality along k 1 = @ 1. The corresponding one-form reads v = h X 2 dx 3 h(1 )X 3 dx 2, 2 R. The constraints for gauging the sigma-model are trivially satisfied. The geometry of the dual background is determined from the quantities G = R 2 1, = d + v, D =0, k = R 2 1 dx 1, M = G = R 2 1. The metric and field strength are obtained as Ǧ = 1 R 2 1 ^? + R 2 2 dx 2 ^?dx 2 + R 2 3 dx 3 ^?dx 3, Ȟ =0.

torus :: one t-duality II As expected, the dual background is a twisted torus (with =1) ďs 2 = 1 R 2 1 d + hx 2 dx 3 2 + R 2 2 dx 2 2 + R 2 3 dx 3 2, Ȟ =0.

torus :: two t-dualities I Consider two collective T-dualities along k 1 = @ 1 and k 2 = @ 2. The constraints on gauging the sigma-model imply (for 2 R ) v 1 = h X 2 dx 3 h(1 )X 3 dx 2, v 2 = h(1 + )X 3 dx 1 + h X 1 dx 3. The geometry of the dual background is determined from R 2 G = 1 0 0 R2 2, = 0 +hx 3 D = hx 3, k 0 = d 1 + v 1, d 2 + v 2 R 2 1 dx 1 R2 2. dx2

torus :: two t-dualities II The metric of the enlarged target space (in the basis {dx 1,dX 2,dX 3, 1, 2 }) reads Ǧ IJ = 1 B @ 0 R1 hx 3 2 0 0 0 0 R1 2 hx 3 R2 hx 3 2 0 +R2 2 hx 3 0 0 0 R3 2 0 0 0 +R2 2 hx 3 0 R2 2 0 R1 2 hx 3 0 0 0 R1 2 1, C A = R 2 1R 2 2 + hx 3 2. Performing then a change of basis one finds T I A = 0 B @ 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 hx 3 0 1 0 +hx 3 0 0 0 1 1 C A Ǧ AB =(T T Ǧ T ) AB = 1 0 B @ 0 0 0 0 0 0 0 0 0 0 0 0 R3 2 0 0 0 0 0 R2 2 0 0 0 0 0 R1 2 1. C A

torus :: two t-dualities III Performing a similar analysis for the field strength and adjusting the notation, one finds ďs 2 = 1 R 2 1 R2 2 + hx 3 2 h R 2 1 d 1 2 + R 2 2 d 2 Ȟ = h R2 1R 2 2 hx 3 2 hr1 2R2 2 + hx 3 i 2 2 d 1 ^ d 2 ^ dx 3. 2 i + R 2 3 dx 3 2, This is the familiar T-fold background.

torus :: three t-dualities Finally, consider three collective T-dualities along k 1 = @ 1, k 2 = @ 2 and k 3 = @ 3. The constraints on gauging the sigma-model require the H-flux to be vanishing k k k H =0 H =0. The dual model for H=0 is characterized by ďs 2 = 1 R 2 1 d 1 2 + 1 R 2 2 d 2 2 + 1 R 2 3 d 3 2, Ȟ =0.

torus :: three t-dualities Finally, consider three collective T-dualities along k 1 = @ 1, k 2 = @ 2 and k 3 = @ 3. The constraints on gauging the sigma-model require the H-flux to be vanishing k k k H =0 H =0. The dual L k[ model v ] = for f H=0 v is characterized by k[ f ďs 2 = 1 2 1 ] v = 1 3 k k k H R1 2 d 1 + R 2 2 d 2 2 + 1 R 2 3 d 3 2, Ȟ =0.

torus :: three t-dualities Finally, consider three collective T-dualities along k 1 = @ 1, k 2 = @ 2 and k 3 = @ 3. The constraints on gauging the sigma-model require the H-flux to be vanishing k k k H =0 H =0. The dual model for H=0 is characterized by ďs 2 = 1 R 2 1 d 1 2 + 1 R 2 2 d 2 2 + 1 R 2 3 d 3 2, Ȟ =0.

torus :: summary The formalism for collective T-duality introduced above works as expected. three-torus with H-flux 1 T-duality twisted torus 2 T-dualities T-fold three-torus with H=0 3 T-dualities torus with R 1/R Answer 1 :: three T-dualities not allowed due to gauging constraint.

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples a) three-torus b) three-sphere 5. summary

sphere :: setting Consider a three-sphere with H-flux, specified by ds 2 = R 2 h sin 2 (d 1 ) 2 + cos 2 (d 2 ) 2 +(d ) 2 i, 1,2 =0...2, H = h 2 2 sin cos d 1 ^ d 2 ^ d, =0... 2. This model is conformal if h =4 2 R 2. The isometry algebra is so(4) = su(2) su(2), and the Killing vectors satisfy [K, K ] L = K, [K, K ] L =0, K 2 = K 2 = R2 4. [ K, K ] L = K,

sphere :: setting Consider a three-sphere with H-flux, specified by 0 h 1 0 i 1 ds 2 = R 2 sin 2 (d 1 ) 2 + cos 2 (d 2 ) 2 +(d ) 2, 1,2 =0...2, K 1 = 1 +1 @ 1 A, K1 = 1 +1 @ +1 A, 2 2 0 0 0 H = h =0... 2 2 sin cos d 1 0 1 1 ^ d 2 ^ d, 2. K 2 = 1 sin( 1 2 ) cot @ sin( 1 2 ) tan A, K2 = 1 +sin( 1 + 2 ) cot @ sin( 1 + 2 ) tan A, 2 2 cos( 1 2 ) cos( 1 + 2 ) This model is conformal if h =4 2 R 2. K 3 = 1 2 0 @ 1 0 cos( 1 2 ) cot cos( 1 2 ) tan A, K3 = 1 @ 2 sin( 1 2 ) + cos( 1 + 2 ) cot cos( 1 + 2 ) tan +sin( 1 + 2 ) 1 A. The isometry algebra is so(4) = su(2) su(2), and the Killing vectors satisfy [K, K ] L = K, [K, K ] L =0, K 2 = K 2 = R2 4. [ K, K ] L = K,

sphere :: setting Consider a three-sphere with H-flux, specified by ds 2 = R 2 h sin 2 (d 1 ) 2 + cos 2 (d 2 ) 2 +(d ) 2 i, 1,2 =0...2, H = h 2 2 sin cos d 1 ^ d 2 ^ d, =0... 2. This model is conformal if h =4 2 R 2. The isometry algebra is so(4) = su(2) su(2), and the Killing vectors satisfy [K, K ] L = K, [K, K ] L =0, K 2 = K 2 = R2 4. [ K, K ] L = K,

sphere :: one t-duality Consider one T-duality along K1. In this case, all constraints are satisfied: constraints from gauging the sigma-model the matrices and are invertible G M X X The dual model is characterized by the metric and H-flux Ǧ = R2 4 h (d ) 2 +sin 2 ( )(d ) 2i + 4 R 2 ^?, Ȟ =sin d ^ d ^, This metric describes a circle fibered over a two-sphere. d = h sin d ^ d. 16 2 Bouwknegt, Evslin, Mathai - 2003 incomplete in Alvarez, Alvarez-Gaume, Barbon, Lozano - 1993

sphere :: two t-dualities I For two collective T-dualities, consider the commuting Killing vectors K 1 and K1. The constraints for this model are almost satisfied: constraints from gauging the sigma-model the matrices and are invertible G M X X The dual model, via the above formalism, takes a form similar to the T-fold Ǧ = R 2 (d ) 2 + 1 (d 1) 2 R 2 sin 2 + h 2 4 2 R 2 cos2 + 1 R 2 cos 2 + (d 2)2 h 4 2 R 2 2 cos4 sin 2, Ȟ = 8h 2 h 2 16 4 R 4 sin cos 16 2 R 4 sin 2 + h 2 cos 2 2 d ^ d 1 ^ d 2.

sphere :: two t-dualities I For two collective T-dualities, consider the commuting Killing vectors K 1 and K1. The constraints for this model are almost satisfied: constraints from gauging the sigma-model the matrices and are invertible G M X X det G = R4 16 sin2 (2 ) The dual model, via the above formalism, takes a form similar to the T-fold Ǧ = R 2 (d ) 2 + 1 (d 1) 2 R 2 sin 2 + h 2 4 2 R 2 cos2 + 1 R 2 cos 2 + (d 2)2 h 4 2 R 2 2 cos4 sin 2, Ȟ = 8h 2 h 2 16 4 R 4 sin cos 16 2 R 4 sin 2 + h 2 cos 2 2 d ^ d 1 ^ d 2.

sphere :: two t-dualities I For two collective T-dualities, consider the commuting Killing vectors K 1 and K1. The constraints for this model are almost satisfied: constraints from gauging the sigma-model the matrices and are invertible G M X X The dual model, via the above formalism, takes a form similar to the T-fold Ǧ = R 2 (d ) 2 + 1 (d 1) 2 R 2 sin 2 + h 2 4 2 R 2 cos2 + 1 R 2 cos 2 + (d 2)2 h 4 2 R 2 2 cos4 sin 2, Ȟ = 8h 2 h 2 16 4 R 4 sin cos 16 2 R 4 sin 2 + h 2 cos 2 2 d ^ d 1 ^ d 2.

sphere :: two t-dualities II But, when starting from a conformal model with h =4 2 R 2, the background becomes G = R 2 (d ) 2 + 1 R 2 h(d 1) 2 + tan 2 (d 2) 2i, H =0. With dual dilaton = log R 2 cos +, this is again a conformal model.

sphere :: three t-dualities Consider finally a non-abelian T-duality along, and. K 1 K 2 K 3 constraints from gauging the sigma-model the matrices G and M are invertible satisfied only for H=0 EP - work in progress The dual model for H=0 is obtained as (with 0 and 1,2 =0,...,2 ) Ǧ = 4 R2 d ^?d + R2 4 2 2 + R4 16 i hd 1 ^?d 1 +sin 2 ( 1 ) d 2 ^?d 2, Ȟ = 2 2 + R4 16 2 apple 2 +3 R4 16 sin( 1) d ^ d 1 ^ d 2. Alvarez, Alvarez-Gaume, Barbon, Lozano - 1993 Curtright, Uematsu, Zachos - 1996

sphere :: summary In the formalism for T-duality introduced above, for a conformal model one finds: three-sphere with H-flux 1 T-duality S 1 fibered over S 2 2 T-dualities non-compact, geometric three-sphere with H=0 3 T-dualities S 2 fibered over a ray Answer 2 :: T-duality for the (conformal) three-sphere leads to geometric spaces.

outline 1. motivation 2. non-geometric spaces 3. collective t-duality 4. examples 5. summary

summary I A formalism of collective T-duality has been developed :: highlighting the underlying structure (enlarged target space), and giving rise to new constraints. Collective T-duality has been studied for two examples :: three-torus three-sphere known results reproduced no non-geometric backgrounds

summary II For two collective T-duality transformations it was found :: three-torus with H-flux (not conformal) 2 T-dualities non-geometric three-sphere with H-flux (conformal) 2 T-dualities geometric Answer 3 :: non-geometry might be related to non-conformality