Nahm equations

From HandWiki

In differential geometry and gauge theory, the Nahm equations are a system of ordinary differential equations introduced by Werner Nahm[1] in the context of the Nahm transform – an alternative to Ward's twistor construction of monopoles. The Nahm equations are formally analogous to the algebraic equations in the ADHM construction of instantons (Nigel Hitchin called them a "bold adaptation"[2]), where finite order matrices are replaced by differential operators.

Deep study of the Nahm equations was carried out by Nigel Hitchin and Simon Donaldson. Conceptually, the equations arise in the process of infinite-dimensional hyperkähler reduction. They can also be viewed as a dimensional reduction of the anti-self-dual Yang-Mills equations.[3] Among their many applications we can mention: Hitchin's construction of monopoles, where this approach is critical for establishing nonsingularity of monopole solutions; Donaldson's description of the moduli space of monopoles; and the existence of hyperkähler structure on coadjoint orbits of complex semisimple Lie groups, proved by Kronheimer,[4] Biquard,[5] and Kovalev.[6]

Equations

Let T1(z),T2(z),T3(z) be three matrix-valued meromorphic functions of a complex variable z. The Nahm equations are a system of matrix differential equations

dT1dz=[T2,T3]dT2dz=[T3,T1]dT3dz=[T1,T2],

together with certain analyticity properties, reality conditions, and boundary conditions. The three equations can be written concisely using the Levi-Civita symbol, in the form

dTidz=12j,kϵijk[Tj,Tk]=j,kϵijkTjTk.

More generally, instead of just considering N×N matrices, one can consider Nahm's equations with values in a Lie algebra 𝔤.

Additional conditions

One often considers the Nahm equations with boundary and reality conditions dependent on the context.

In the original context of SU(2) monopoles the variable z is restricted to the open interval (0,2), and the following conditions are imposed:

  1. Ti*=Ti;
  2. Ti(2z)=Ti(z)T;
  3. Ti can be continued to a meromorphic function of z in a neighborhood of the closed interval [0,2], analytic outside of 0 and 2, and with simple poles at z=0 and z=2; and
  4. At the poles, the residues of T1,T2,T3 form the N-dimensional irreducible representation of the group SU(2).

Nahm–Hitchin description of monopoles

There is a natural equivalence between

  1. the monopoles of charge K for the group SU(2), modulo gauge transformations, and
  2. the solutions of Nahm equations satisfying the additional conditions above, modulo the simultaneous conjugation of T1,T2,T3 by the group O(k,R).

Lax representation

The Nahm equations can be written in the Lax form as follows. Set

A0=T1+iT2,A1=2iT3,A2=T1iT2[3pt]A(ζ)=A0+ζA1+ζ2A2,B(ζ)=12dAdζ=12A1+ζA2,

then the system of Nahm equations is equivalent to the Lax equation

dAdz=[A,B].

As an immediate corollary, we obtain that the spectrum of the matrix A does not depend on z. Therefore, the characteristic equation

det(λI+A(ζ,z))=0,

which determines the so-called spectral curve in the twistor space T1 is invariant under the flow in z. Moreover, the flow of the ODE linearises on the Jacobian variety of the spectral curve [7]

See also

  • Bogomolny equation
  • Yang–Mills–Higgs equations

References

  1. (Nahm 1981)
  2. (Hitchin 1983)
  3. (Donaldson 1984)
  4. (Kronheimer 1990)
  5. (Biquard 1996)
  6. (Kovalev 1996)
  7. (Griffiths 1985)