Topology in Condensed Matter Physics

January 27, 2009 by

During the spring semester there will be a course from the physics department with the above title. You can find more info here. I went today to the first lecture and it looks good.

Posted in Announcement | Leave a Comment »

Yang-Mills Instantons (I)

October 15, 2008 by

In this series of post I would like to scratch the surface of an enormous iceberg called “instanton physics”. First I would like to mention some references:

  1. arXiv:0802.1862 – Lectures on instantons by Vandoren and van Nieuwenhuizen,
  2. arXiv:hep-th/0206063 – The calculus of many instantons by Dorey, Hollowood, Khoze and Mattis,
  3. arXiv:hep-th/0004186 – Yang-Mills- and D-instantons by Belitsky, Vandoren and van Nieuwenhuizen.

Most of these posts are going to be very loyal to the first item, meaning that I will only discuss one instanton cases. As the title suggests, the second item deals with the case of many instantons.

So what is an instanton? A Yang-Mills instanton is a solution to the classical field equations in Euclidean space that give a finite action. Next you might ask, why finite action? Recall that for (classical) Yang-Mills the path integral has the form

Z_{E}\left(A\right) = \displaystyle\int DA \exp{\left(-\frac{S_{cl}}{\hbar}\right)}.

A finite action then gives the leading contribution to Z_{E}\left(A\right). We can distinguish between regular instantons, ones that have a singularity at Euclidean infinity, and singular instanton, which don’t have singularity at infinity but at some point in space x_{0}^{m}. It turns out that we can map singular to regular solutions by a singular gauge transformation.

Later we will consider systems in the background of an instanton. We can achive this by the usual ways of minimal coupling,

\partial \rightarrow \partial + A .

When we have such a background one has to be careful with zero modes. These are solutions of the linearized field equations that are normalizable. Alternatively, zero modes are eigenfunctions of the quantum operator with zero eigenvalue. The quantum operator is (I think) the operator that appears in the action when one integrates by part the Lagrangian. For example,

S = \int d^4 x \partial X \partial X \rightarrow -\int d^4 x X \partial^{2} X.

In this case the quantum operator corresponds to \partial^2. Zero modes have their own measure in the path integral and sometimes they are the only contribution (e.g. in supersymmetric theories the non-zero modes cancel).

Let us be a bit precise. Let us consider Yang-Mills gauge theory in 4 Euclidean space dimensions with gauge group SU(N). The Lie algebra has generators T_{a} that are traceless, anti-hermitian N \times N matrices with the normalization

tr\left(T_{a}T_{b}\right) = -\displaystyle\frac{1}{2}\delta_{ab}.

The action for Yang-Mills theory is

S = \displaystyle\frac{-1}{2 g^{2}} \int_{\mathbb{R}^{4}} d^{4}x \left[ tr\left(F^{mn}F_{mn}\right) \right],

with the field strength given by

F_{mn} = \partial_{m} A_{n} - \partial_{n}A_{m} + \left[A_{m} , A_{n}\right].

The classical field equations for A_{m} are found from the Euler-Lagrange equations:

D_{m} F^{mn} = 0 \qquad D_{m} \cdot = \partial_{m} \cdot + \left[A_{m}, \cdot \right] .

Since instantons are solutions to this equation but have finite action, we expect the field strength to vanish very far away from the origin. Since F appears quadratic in the action, it should vanish faster than r^{2}. The statement that the field strength vanishes leads us to looking for gauge potentials that are pure gauge, that is, they have the form

A_{m} = U^{-1} \partial_{m} U \qquad U \in SU(N).

[To be continued…]

Tags:
Posted in Gauge theory | 1 Comment »

Seminar returns

October 11, 2008 by

Last week saw the first seminar of this fall semester. We are meeting on Wednesdays, 2:15 PM at Eitan’s office (Math 2-122).

This week’s seminar (October 15th) will be about Instantons from the physics side.

Posted in Organizational | 2 Comments »

Return from the summer

September 12, 2008 by

Exactly a week ago I took my oral exam and passed, and soon it will be Eitan’s turn. This explains why things have been a bit quite around here. The bad news is that things will remain quite for a bit longer.

The good news is that we are planning to start again the meetings during the fall with topological field theory! I know nothing about this, erm, field. It will take some time to read and prepare something. But I already started to look into some notes on the arXiv:

I will probably take care of the physics part, which is mostly included in the second link. Hopefully we can do some quantum stuff.

Posted in Organizational | 1 Comment »

Weak Interactions

August 13, 2008 by

I found a rather nice article by Witten on the weak interactions and gauge symmetry breaking. It makes use of the terminology Eitan and Brandon have been presenting. The article can be found here.

Posted in Gauge theory, References | Leave a Comment »

Week 07 seminar

August 12, 2008 by

Anybody has problems with the seminar being tomorrow (Wednesday) at 3:30 PM? The topic will be Electroweak and a bit of Yang-Mills.

Posted in Organizational | Leave a Comment »

Electrodynamics on a principal bundle IV

August 7, 2008 by

Consider the matrix group O(1,3), i.e. matrices B such that B^{T}\eta B=\eta where \eta=diag(1,-1,-1,-1), or equivalently \eta(Bv,Bw)=\eta(v,w) for any events v,w in Minkowski spacetime. This group has 4 connected components coming from det(B)=\pm1 and B_{00}>0 or B_{00}<0. The component containing the identity is called the proper, orthochronous Lorentz group L=L_{+}^{\uparrow}. Physically it contains all rotations, and boosts (Lorentz tranformations) and so dim(L)=6.

We can cover L by the simply connected group SL(2,\mathbb{C}), i.e. 2\times2 complex matrices A with det(A)=1. First we identify Minkowski spacetime \mathbb{R}^{4} with the space of 2\times2 Hermitian matrices, i.e. matrices H such that \overline{H}^{T}=H, in such a way that if H is the Hermitian matrix identified with the event x then det(H)=|x|^{2}. Then we can define a covering map \Lambda:SL(2,\mathbb{C})\to L by identifying \Lambda(A)x with AH\overline{A}^{T}. We have that \Lambda(A)\in L since
|\Lambda(A)x|^{2}=det(AH\overline{A}^{T})=det(A)det(H)det(A)=det(H)=|x|^{2}. It can be shown that \Lambda is a 2-1 homomorphism of Lie groups.

Now, there are two important irreducible representations for SL(2,\mathbb{C}) on \mathbb{C}^{2}, the “spin \frac{1}{2}” representations given by multiplication A\left(\begin{array}{c} z_{1}\\ z_{2}\end{array}\right) and multiplication by the adjoint \overline{A}^{T}\left(\begin{array}{c} z_{1}\\z_{2}\end{array}\right). The Dirac representation is the direct sum of these representations \left(\begin{array}{cc}A& 0\\ 0&\overline{A}^{T}\end{array}\right) \left(\begin{array}{c}z_{1}\\z_{2}\\z_{3}\\z_{4}\end{array}\right).

Let \pi:FM\to M be the orthonormal frame bundle for spacetime. Its fibers F_{m}M are ordered orthonormal bases of T_{m}M, or equivalently isometries p:\mathbb{R}^{4}\to T_{m}M. There is a right action of O(1,3) given by right composition pB which makes the frame bundle an O(1,3)-bundle. We say that M is space and time orientable iff FM has 4 components and a choice of component FM_{0} is a space and time orientation. Then the restriction \pi:FM_{0}\to M is an L-bundle.

The solder form is an \mathbb{R}^{4}-valued 1-form \phi on FM_{0} given by \phi_{p}(X)=p^{-1}(\pi_{*}(X)). The torsion of a connection \theta on FM_{0} is \Theta=d\phi+\theta\wedge\phi. It turns out that there is a unique connection whose torsion is \Theta=0. This is the Levi-Civita connection \theta.

A spin structure on M is a manifold SM and a smooth map \lambda:SM\to FM_{0} such that \pi\circ\lambda:SM\to M is an SL(2,\mathbb{C})-bundle with \lambda(pA)=\lambda(p)\Lambda(A). We can define a connection \tilde{\theta} on SM by \tilde{\theta}=\Lambda_{*}^{-1}\lambda^{*}\theta where \Lambda_{*} is the isomorphism of Lie algebras induced by \Lambda:SL(2,\mathbb{C})\to L.

Now consider sections \psi of the vector bundle associated to SM by the Dirac representation. Dirac’s idea was to introduce an operator \not\hspace{-4pt}D such that \not\hspace{-4pt}D^{2}=\square, i.e. the Dirac operator is the “square root” of the d’Alembert operator. A full understanding of the Dirac operator requires Clifford algebras, i.e the algebra generated over Minkowski space modulo v^{2}=\eta(v,v). It turns out that the smallest representation \gamma of this Clifford algebra is 4-dimensional which is why we need a 4-dimensional representation of SL(2,\mathbb{C}) as well. Then we can define the Dirac operator as \not\hspace{-4pt}D=\eta(\gamma,\nabla) where \nabla is the connection associated to \tilde{\theta} and we inner product them somehow.

In more detail for the d’Alembertian on Minkowski spacetime, \square=\frac{\partial^{2}}{\partial t^{2}}-\frac{\partial^{2}}{\partial x^{2}}-\frac{\partial^{2}}{\partial y^{2}}-\frac{\partial^{2}}{\partial z^{2}}, define

\not\hspace{-4pt}D=\left(\begin{array}{cccc} 1& 0& 0& 0\\ 0 & 1& 0& 0\\ 0 & 0& -1& 0\\ 0 & 0& 0& -1\end{array}\right)\frac{\partial}{\partial t}+\left(\begin{array}{cccc} 0& 0& 0& 1\\ 0 & 0& 1& 0\\ 0 & -1& 0& 0\\ -1& 0& 0& 0\end{array}\right)\frac{\partial}{\partial x}

+\left(\begin{array}{cccc} 0& 0& 0& -i\\ 0 & 0& i& 0\\ 0 & i& 0& 0\\ -i& 0& 0& 0\end{array}\right)\frac{\partial}{\partial y}+\left(\begin{array}{cccc}0& 0& 1& 0\\ 0 & 0& 0& -1\\ -1& 0& 0& 0\\ 0 & 1& 0& 0\end{array}\right)\frac{\partial}{\partial z}

We can work out that \not\hspace{-4pt}D^{2}=\square.

Then we demand that the Dirac equation holds, \not\hspace{-4pt}D\psi=m\psi. This gives us a theory of a spin-\frac{1}{2} particle, an electron or positron, but we have not yet coupled it to electromagnetism.

We can splice a G_{1}-bundle \pi_{1}:P_{1}\to M with a G_{2}-bundle \pi_{2}:P_{2}\to M. Define P=\{(p_{1},p_{2})\in P_{1}\times P_{2}:\pi_{1}(p_{1})=\pi_{2}(p_{2})\} and \pi:P\to M by \pi(p_{1},p_{2})=\pi_{1}(p_{1})=\pi_{2}(p_{2}). This is a G_{1}\times G_{2}-bundle with (p_{1},p_{2})(g_{1},g_{2})=(p_{1}g_{1},p_{2}g_{2}). Given connections \omega_{1},\omega_{2} on P_{1},P_{2}, we can define a connection \omega on P by \omega=\pi^{1*}\omega_{1}\oplus\pi^{2*}\omega_{2} with \pi^{i}:P\to P_{i} given by \pi^{i}(p_{1},p_{2})=p_{i}.

Splice together our U(1)-bundle P with SM and also splice \omega with \tilde{\theta}. Consider the representation of U(1)\times SL(2,\mathbb{C}) on \mathbb{C}^{4} given by combining the Dirac representation with multiplication by e^{i\lambda}. We get an associated vector bundle with an associated connection and Dirac operator \not\hspace{-4pt}D. A charged electron coupled to electromagnetism is then a section \psi for which the Dirac equation \not\hspace{-4pt}D\psi=m\psi holds.

Posted in Gauge theory, Geometry | 4 Comments »

Electrodynamics on a principal bundle III

August 4, 2008 by

Suppose we had a principal U(1)-bundle \pi:P\to M with a connection \omega with curvature \Omega.

The Lie algebra \mathfrak{u}(1) is just the set of imaginary numbers i\mathbb{R} with trivial Lie bracket {[},{]}=0. The local potential is a real-valued 1-form A_{U} defined by \omega_{U}=iA_{U}. The local field strength F_{U} is defined by \Omega_{U}=iF_{U}.

A change of gauge is given by g_{UV}=e^{i\lambda} with \lambda:U\cap V\to\mathbb{R}. We see that local connections are related by \omega_{V}=e^{-i\lambda}\omega_{U}e^{i\lambda}+e^{-i\lambda}de^{i\lambda}=\omega_{U}+id\lambda, so that local potentials are related by A_{V}=A_{U}+d\lambda. Local curvatures are related by \Omega_{V}=e^{-i\lambda}\Omega_{U}e^{i\lambda}=\Omega_{U}, so that local field strengths are related by F_{U}=F_{V}. This means that the field strength is globally defined on M.

By the Bianchi identity we have d\Omega=0 so dF=0, so the homogeneous Maxwell equation comes along for free. We can get the inhomogeneous Maxwell equation by requiring that d*F=*J.

Now, consider the action U(1) on \mathbb{C} given by multiplication e^{i\lambda}z. Associated to our principal U(1) bundle we get a vector bundle with fiber \mathbb{C} with an induced connection \nabla locally given by \nabla=d+\omega_U=d+iA_U. We will write sections of the associated bundle as \psi. We can define the d’Alembert operator \square=*\nabla*\nabla+\nabla*\nabla*. If we require the Klein-Gordon equation, \square\psi=m^2\psi, then we have a theory of a charged spin-0 particle coupled to electromagnetism.

In order to couple electromagnetism to more interesting particles like Dirac’s electron, we need to incorporate spin somehow.

Posted in Gauge theory, Geometry | Leave a Comment »

Electrodynamics on a principal bundle II

August 4, 2008 by

We will assume G is a group of matrices. A principal G-bundle is a smooth surjection of manifolds \pi:P\to M with a free transitive right action R of G on P such that \pi^{-1}\pi(p)=pG and for any m\in M there is an open set U with m\in U\subset M and a diffeomorphism T_U=\pi\times t_U:\pi^{-1}(U)\to U\times G called a “local trivialization” such that t_U(pg)=t_U(p)g. Local trivializations correspond to the physical notion of “choice of gauge”.

Intuitively, P is a manifold composed of copies of the group G parametrized by the base space M. A good example is the boundary of the Mobius strip which can be thought of as a \mathbb{Z}/2-bundle over S^1.

A useful notion is that of a local section \sigma_U:U\to P with U an open set with U\subset M such that \pi(\sigma_U(m))=m. It can be shown that there is a canonical 1-1 correspondence between local sections \sigma_U and local trivializations T_U.

Define transition functions g_{UV}:U\cap V\to G by g_{UV}(m)=t_U(p)t_V(p)^{-1} where \pi(p)=m. This is well defined since t_U(pg)t_V(pg)^{-1}=t_U(p)gg^{-1}t_V(p)^{-1}=t_U(p)t_V(p)^{-1}. Transition functions correspond to the physical notion of “change of gauge”. We can relate any two local sections by \sigma_V=\sigma_U g_{UV}.

Let \mathfrak{g} be the Lie algebra for G. A connection \omega is a \mathfrak{g}-valued 1-form on P such that if If X\in\mathfrak{g} and \tilde{X} is the tangent field on P given by \tilde{X}_{p}=\frac{d}{dt}pe^{tX}|_{t=0}, then \omega(\tilde{X})=X. Also we require that R(g)^{*}(\omega)=g^{-1}\omega g.

We define local connections on M by \omega_U=\sigma_U^*\omega. Local connections are related by \omega_{V}=g_{UV}^{-1}\omega_{U}g_{UV}+g_{UV}^{-1}dg_{UV}.

We define curvature \Omega=d\omega+\frac{1}{2}[\omega,\omega] meaning \Omega(X,Y)=d\omega(X,Y)+\frac{1}{2}[\omega(X),\omega(Y)]. We can define local curvature by \Omega_U=\sigma_U^*\Omega. Local curvatures are then related by \Omega_V=g_{UV}^{-1}\Omega_U g_{UV}. The Bianchi identity says d\Omega=[\omega,\Omega].

We are now in a position to define electrodynamics on a principal bundle.

Posted in Gauge theory, Geometry | 2 Comments »

Electrodynamics on a principal bundle I

August 4, 2008 by

Maxwell’s equations in relativistically covariant form are

\partial_{\mu}F^{\mu\nu}=J^{\nu}
\partial_{[\lambda}F_{\mu\nu]}=0

Since F_{\mu\nu}=-F_{\nu\mu} we can define a 2-form F=F_{\mu\nu}dx^{\mu}dx^{\nu}. We can also define a 1-form J=J_\mu dx^\mu. Then we can re-express Maxwell’s equations using exterior differentiation and the Hodge star.

d*F=*J
dF=0

The continuity equation d*J=0 then follows from the inhomogeneous Maxwell equation. We expect from the homogeneous Maxwell equation that F=dA. In fact this is only true locally. This means that for every event m in our spacetime M there is an open set U with m\in U\subset M and a 1-form A_U on U with F|_U=dA_U. This follows from Poincare’s lemma.

We cannot say the A exists globally. For instance if F=sin\phi d\phi d\theta, the area form of the unit sphere in spherical coordinates, then dF=cos\phi d\phi d\phi d\theta=0 since d\phi d\phi=0 by antisymmetry of wedge product of 1-forms. Also, taking \Sigma to be the unit sphere, we know that \int_\Sigma F=4\pi. However, by Stokes’ Theorem, if F=dA then \int_\Sigma F=\int_\Sigma dA=\int_{\partial \Sigma} A=0\neq 4\pi. So, we cannot have F=dA globally.

Physically we interpret this as a magnetic monopole with magnetic charge 4\pi and worldline, the time axis, r=0. Mathematically, what is happening is that the complement of the time axis has nontrivial topology. Specifically its second de Rham cohomology is nontrivial. Intuitively, there is a kind of 2-dimensional, spherical “hole” in the complement of the time axis.

In addition to being nonglobal, the potential A is defined only up to addition of a closed 1-form since d(A+d\lambda)=dA=F. We would like to find a global mathematical object corresponding to the potential which doesn’t depend on our “choice of gauge”. This is our motivation for understanding connections on principal bundles.

Posted in Gauge theory, Geometry | Leave a Comment »

« Older Entries
Design a site like this with WordPress.com
Get started