In differential geometry and gauge theory, the Nahm equations are a system of ordinary differential equations introduced by Werner Nahm 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, 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 (Donaldson 1984). 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 1990), (Biquard 1996), and (Kovalev 1996).

Equations

Let ${\displaystyle T_{1}(z),T_{2}(z),T_{3}(z)}$ be three matrix-valued meromorphic functions of a complex variable ${\displaystyle z}$. The Nahm equations are a system of matrix differential equations

${\displaystyle {\begin{aligned}{\frac {dT_{1)){dz))&=[T_{2},T_{3}]\\[3pt]{\frac {dT_{2)){dz))&=[T_{3},T_{1}]\\[3pt]{\frac {dT_{3)){dz))&=[T_{1},T_{2}],\end{aligned))}$

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

${\displaystyle {\frac {dT_{i)){dz))={\frac {1}{2))\sum _{j,k}\epsilon _{ijk}[T_{j},T_{k}]=\sum _{j,k}\epsilon _{ijk}T_{j}T_{k}.}$

More generally, instead of considering ${\displaystyle N}$ by ${\displaystyle N}$ matrices, one can consider Nahm's equations with values in a Lie algebra ${\displaystyle g}$.

Additional conditions

The variable ${\displaystyle z}$ is restricted to the open interval ${\displaystyle (0,2)}$, and the following conditions are imposed:

1. ${\displaystyle T_{i}^{*}=-T_{i};}$
2. ${\displaystyle T_{i}(2-z)=T_{i}(z)^{T};\,}$
3. ${\displaystyle T_{i}N}$ can be continued to a meromorphic function of ${\displaystyle z}$ in a neighborhood of the closed interval ${\displaystyle [0,2]}$, analytic outside of ${\displaystyle 0}$ and ${\displaystyle 2}$, and with simple poles at ${\displaystyle z=0}$ and ${\displaystyle z=2}$; and
4. At the poles, the residues of ${\displaystyle T_{1},T_{2},T_{3))$ form an irreducible representation of the group SU(2).

Nahm–Hitchin description of monopoles

There is a natural equivalence between

1. the monopoles of charge ${\displaystyle K}$ for the group ${\displaystyle SU(2)}$, modulo gauge transformations, and
2. the solutions of Nahm equations satisfying the additional conditions above, modulo the simultaneous conjugation of ${\displaystyle T_{1},T_{2},T_{3))$ by the group ${\displaystyle O(k,R)}$.

Lax representation

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

${\displaystyle {\begin{aligned}&A_{0}=T_{1}+iT_{2},\quad A_{1}=-2iT_{3},\quad A_{2}=T_{1}-iT_{2}\\[3pt]&A(\zeta )=A_{0}+\zeta A_{1}+\zeta ^{2}A_{2},\quad B(\zeta )={\frac {1}{2)){\frac {dA}{d\zeta ))={\frac {1}{2))A_{1}+\zeta A_{2},\end{aligned))}$

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

${\displaystyle {\frac {dA}{dz))=[A,B].}$

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

${\displaystyle \det(\lambda I+A(\zeta ,z))=0,}$

which determines the so-called spectral curve in the twistor space ${\displaystyle TP^{1))$is invariant under the flow in ${\displaystyle z}$.