Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
Making sense of nonHermitian Hamiltonians
Crab LenderPractised NymphetsWashing Nervy Tuitions
Making sense of nonHermitian Hamiltonians
Carl BenderPhysics DepartmentWashington University
Quantum Mechanics
• “Anyone who thinks he can contemplate quantum mechanics without getting dizzy hasn’t properly understood it.” – Niels Bohr
• “Anyone who thinks they know quantum mechanics doesn’t.” – Richard Feynman
• “I don’t like it, and I’m sorry I ever had anything to do with it.” – Erwin Schrödinger
Axioms of Quantum Mechanics
• Causality• Locality• Stability of the vacuum• Relativity• Probabilistic interpretation
Dirac Hermiticity …
• guarantees real energy and conserved probability• but … is mathematical and not physical
32 ixpH +=
Wait a minute…This model has
PT symmetry
32 ixpH +=
Some references …
• CMB and S. Boettcher, PRL 80, 5243 (1998)• CMB, D. Brody, H. Jones, PRL 89, 270401 (2002)• CMB, D. Brody, and H. Jones, PRL 93, 251601 (2004) • CMB, D. Brody, H. Jones, B. Meister, PRL 98, 040403 (2007)• CMB and P. Mannheim, PRL 100, 110402 (2008)• CMB, Reports on Progress in Physics 70, 947 (2007)• P. Dorey, C. Dunning, and R. Tateo, JPA 34, 5679 (2001)• P. Dorey, C. Dunning, and R. Tateo, JPA 40, R205 (2007)
The originaldiscoverers of PT symmetry:
MY TALK
Outline of Talk
• Beginning• Middle• End• (applause)
How to “prove” that the eigenvalues are real
PT Boundary
How to “prove” that the eigenvalues are real
The proof is really hard!You need to use (3)Bethe ansatz(4)Monodromy group(5)Baxter TQ relation(6)Functional Determinants
PT BoundaryGreatest murder mystery of all time…Extinction of over 90% of species!
Permian era Triassic era
PT Boundary
OK, so the eigenvalues are real …But is this quantum mechanics??
• Probabilistic interpretation??• Hilbert space with a positive metric??• Unitarity??
Dirac, Bakerian Lecture 1941,Proceedings of the Royal Society A
The Hamiltonian determines its own adjoint
Unitarity
With respect to the CPT adjointthe theory has UNITARY timeevolution.
Norms are strictly positive!Probability is conserved!
OK, we have unitarity…But is PT quantum mechanics useful??
• It revives quantum theories that were thought to be dead
• It is beginning to be observed experimentally
The Lee Model
The problem:
“A nonHermitian Hamiltonian is unacceptable partly because it may lead to complex energy eigenvalues, but chiefly because it implies a nonunitary S matrix, which fails to conserve probability and makes a hash of the physical interpretation.”
PT quantum mechanics to the rescue…
Meep! Meep!
PT
GHOSTBUSTING: Reviving quantum theories that were thought to be dead
Gives a fourthorder field equation:
CMB and P. Mannheim, Phys. Rev. Lett. 100, 110402 (2008)CMB and P. Mannheim, Phys. Rev. D 78, 025002 (2008)
PaisUhlenbeck action
The problem: A fourthorder fieldequation gives a propagator like
GHOST!
There are now two possible realizations…
There can be many realizations!
Equivalent Dirac Hermitian Hamiltonian:
Noghost theorem for the fourthorder derivative PaisUhlenbeckmodel, CMB and P. Mannheim, PRL 100, 110402 (2008)
TOTALITARIAN PRINCIPLE
“Everything which is not forbidden is compulsory.” M. GellMann
And there are now observations in tabletop optics experiments!
• Z. Musslimani, K. Makris, R. ElGanainy, and D. Christodoulides, PRL 100, 030402 (2008)
• K. Makris, R. ElGanainy, D. Christodoulides, and Z. Musslimani, PRL 100, 103904 (2008)
Observing PT symmetry using wave guides:
Date: Thu, 13 Mar 2008 23:04:45 0400From: Demetrios Christodoulides <[email protected]>To: Carl M. Bender <[email protected]>Subject: Re: Benasque workshop on nonHermitian Hamiltonians
Dear Carl,I have some good news from Greg Salamo (U. of Arkansas). His students (who are nowvisiting us here in Florida) have just observed a PT phase transition in a passive AlGaAswaveguide system. We will be submitting soon these results as a postdeadline paper toCLEO/QELS and subsequently to a regular journal. We are still fighting against theKramersKronig relations, but the phase transition effect is definitely there. We expect evenbetter results under TE polarization conditions. I will bring them over to Israel.
In close collaboration with us, more teams (also best friends!) are moving ahead in thisdirection. Moti Segev (from Technion) is planning an experiment in an activepassive dualcore optical fiber fabricated in Southampton, England. More experiments will be carriedlater in Germany by Detlef Kip. Christian (his post doc) just left from here with a possibledesign. If everything goes well, with a bit of luck we may have an experimental explosionin the PT area. I wish the funding situation was a bit better. So far everything is done on ashoestring budget (it is subsidized by other projects). Let us see...
All the bestDemetri
OK, but how do we interpret a nonHermitian Hamiltonian??
Solve the quantum brachistochrone problem…
Classical Brachistochrone
• Newton• Bernoulli • Leibniz • L'Hôpital
Classical Brachistochroneis a cycloid
Gravitationalfield
Quantum Brachistochrone
Constraint:
Hermitian case
becomes
Minimize t over all positive rwhile maintaining constraint
Minimum evolution time:
Looks like uncertainty principle but is merelyrate times time = distance
NonHermitian PTsymmetric Hamiltonian
where
Exponentiate H
Interpretation…
Finding the optimal PTsymmetricHamiltonian amounts to constructing a wormhole in Hilbert space!
“The shortest path between two truths in the real domain passes through the complex domain.”
Jacques Hadamard[The Mathematical
Intelligencer 13 (1991)]
Overviewof talk:
Classical PT symmetry
Provides an intuitive explanation of what is going on…
Motion on the Real Axis
Motion of particles is governed by Newton’s Law:
F=maIn freshman physics this motion is restricted to theREAL AXIS.
Harmonic Oscillator: Particle on a Spring
Turning point Turning point
Back and forth motionon the real axis:
( = 0)
Hamilton’s equations
Harmonic Oscillator:
Turning point Turning point
Motion in thecomplex plane:
( = 0)
32 ixpH += ( = 1)
Broken PT symmetry
(ε < 0)
= – 2 11 sheets
The Beast = 16/15
LotkaVolterra equations
Other PTsymmetric classical systems
• KdV equation• CamassaHolm equation• SineGordon equation• Boussinesq equation
1
Making sense of nonHermitian Hamiltonians
Crab LenderPractised NymphetsWashing Nervy Tuitions
2
Making sense of nonHermitian Hamiltonians
Carl BenderPhysics DepartmentWashington University
3
Quantum Mechanics
• “Anyone who thinks he can contemplate quantum mechanics without getting dizzy hasn’t properly understood it.” – Niels Bohr
• “Anyone who thinks they know quantum mechanics doesn’t.” – Richard Feynman
• “I don’t like it, and I’m sorry I ever had anything to do with it.” – Erwin Schrödinger
4
Axioms of Quantum Mechanics
• Causality• Locality• Stability of the vacuum• Relativity• Probabilistic interpretation
5
Dirac Hermiticity …
• guarantees real energy and conserved probability• but … is mathematical and not physical
6
32 ixpH +=
7
Wait a minute…This model has
PT symmetry
32 ixpH +=
8
9
Some references …
• CMB and S. Boettcher, PRL 80, 5243 (1998)• CMB, D. Brody, H. Jones, PRL 89, 270401 (2002)• CMB, D. Brody, and H. Jones, PRL 93, 251601 (2004) • CMB, D. Brody, H. Jones, B. Meister, PRL 98, 040403 (2007)• CMB and P. Mannheim, PRL 100, 110402 (2008)• CMB, Reports on Progress in Physics 70, 947 (2007)• P. Dorey, C. Dunning, and R. Tateo, JPA 34, 5679 (2001)• P. Dorey, C. Dunning, and R. Tateo, JPA 40, R205 (2007)
10
The originaldiscoverers of PT symmetry:
11
12
13
14
15
16
17
18
19
MY TALK
20
Outline of Talk
• Beginning• Middle• End• (applause)
21
How to “prove” that the eigenvalues are real
22PT Boundary
23
How to “prove” that the eigenvalues are real
The proof is really hard!You need to use (3)Bethe ansatz(4)Monodromy group(5)Baxter TQ relation(6)Functional Determinants
24
PT BoundaryGreatest murder mystery of all time…Extinction of over 90% of species!
Permian era Triassic era
PT Boundary
25
OK, so the eigenvalues are real …But is this quantum mechanics??
• Probabilistic interpretation??• Hilbert space with a positive metric??• Unitarity??
26
Dirac, Bakerian Lecture 1941,Proceedings of the Royal Society A
27
The Hamiltonian determines its own adjoint
28
Unitarity
With respect to the CPT adjointthe theory has UNITARY timeevolution.
Norms are strictly positive!Probability is conserved!
29
OK, we have unitarity…But is PT quantum mechanics useful??
• It revives quantum theories that were thought to be dead
• It is beginning to be observed experimentally
30
The Lee Model
31
The problem:
32
33
“A nonHermitian Hamiltonian is unacceptable partly because it may lead to complex energy eigenvalues, but chiefly because it implies a nonunitary S matrix, which fails to conserve probability and makes a hash of the physical interpretation.”
34
PT quantum mechanics to the rescue…
Meep! Meep!
PT
35
GHOSTBUSTING: Reviving quantum theories that were thought to be dead
36
Gives a fourthorder field equation:
CMB and P. Mannheim, Phys. Rev. Lett. 100, 110402 (2008)CMB and P. Mannheim, Phys. Rev. D 78, 025002 (2008)
PaisUhlenbeck action
37
The problem: A fourthorder fieldequation gives a propagator like
GHOST!
38
There are now two possible realizations…
39
There can be many realizations!
Equivalent Dirac Hermitian Hamiltonian:
40
Noghost theorem for the fourthorder derivative PaisUhlenbeckmodel, CMB and P. Mannheim, PRL 100, 110402 (2008)
41
TOTALITARIAN PRINCIPLE
“Everything which is not forbidden is compulsory.” M. GellMann
42
And there are now observations in tabletop optics experiments!
• Z. Musslimani, K. Makris, R. ElGanainy, and D. Christodoulides, PRL 100, 030402 (2008)
• K. Makris, R. ElGanainy, D. Christodoulides, and Z. Musslimani, PRL 100, 103904 (2008)
Observing PT symmetry using wave guides:
43
Date: Thu, 13 Mar 2008 23:04:45 0400From: Demetrios Christodoulides <[email protected]>To: Carl M. Bender <[email protected]>Subject: Re: Benasque workshop on nonHermitian Hamiltonians
Dear Carl,I have some good news from Greg Salamo (U. of Arkansas). His students (who are nowvisiting us here in Florida) have just observed a PT phase transition in a passive AlGaAswaveguide system. We will be submitting soon these results as a postdeadline paper toCLEO/QELS and subsequently to a regular journal. We are still fighting against theKramersKronig relations, but the phase transition effect is definitely there. We expect evenbetter results under TE polarization conditions. I will bring them over to Israel.
In close collaboration with us, more teams (also best friends!) are moving ahead in thisdirection. Moti Segev (from Technion) is planning an experiment in an activepassive dualcore optical fiber fabricated in Southampton, England. More experiments will be carriedlater in Germany by Detlef Kip. Christian (his post doc) just left from here with a possibledesign. If everything goes well, with a bit of luck we may have an experimental explosionin the PT area. I wish the funding situation was a bit better. So far everything is done on ashoestring budget (it is subsidized by other projects). Let us see...
All the bestDemetri
44
45
46
47
48
49
OK, but how do we interpret a nonHermitian Hamiltonian??
Solve the quantum brachistochrone problem…
50
Classical Brachistochrone
• Newton• Bernoulli • Leibniz • L'Hôpital
51
Classical Brachistochroneis a cycloid
Gravitationalfield
52
Quantum Brachistochrone
Constraint:
53
Hermitian case
54
becomes
55
Minimize t over all positive rwhile maintaining constraint
Minimum evolution time:
Looks like uncertainty principle but is merelyrate times time = distance
56
NonHermitian PTsymmetric Hamiltonian
where
57
Exponentiate H
58
Interpretation…
Finding the optimal PTsymmetricHamiltonian amounts to constructing a wormhole in Hilbert space!
59
“The shortest path between two truths in the real domain passes through the complex domain.”
Jacques Hadamard[The Mathematical
Intelligencer 13 (1991)]
60
Overviewof talk:
61
Classical PT symmetry
Provides an intuitive explanation of what is going on…
62
Motion on the Real Axis
Motion of particles is governed by Newton’s Law:
F=maIn freshman physics this motion is restricted to theREAL AXIS.
63
Harmonic Oscillator: Particle on a Spring
Turning point Turning point
Back and forth motionon the real axis:
( = 0)
64
Hamilton’s equations
65
Harmonic Oscillator:
Turning point Turning point
Motion in thecomplex plane:
( = 0)
66
32 ixpH += ( = 1)
67
68
Broken PT symmetry
(ε < 0)
69= – 2 11 sheets
70
71
72
73
74
75
76
The Beast = 16/15
77
LotkaVolterra equations
78
79
80
81
Other PTsymmetric classical systems
• KdV equation• CamassaHolm equation• SineGordon equation• Boussinesq equation