The Nicolai map and stochastic quantisation of supersymmetric theories

Intuition Beginner

Some quantum theories carry a hidden symmetry called supersymmetry that pairs up bosonic and fermionic pieces. The Nicolai map is a change of variables that exploits this pairing to make the theory look as simple as possible. You rename the bosonic field by a new variable, chosen with care, and after the renaming the bosonic part of the action looks like a plain bell-curve weight — the simplest probability distribution there is.

The renaming does two jobs at once. First, it flattens the bosonic landscape into a free Gaussian. Second, the bookkeeping factor that comes from any change of variables — the stretching factor of the rename — turns out to exactly match the factor the fermions contribute. The two cancel. The whole theory collapses into a free Gaussian integral that anyone can do.

Here is the surprise. The new variable is not abstract: it is the random kick of a noisy-ball simulation. Recall the recipe where a ball rolls downhill on a landscape and gets random kicks, settling into a distribution. The Nicolai variable is precisely that random kick, and the downhill direction is set by a function called the superpotential. So a supersymmetric theory is the same thing as a noisy-ball recipe with a clean change of variables for the noise. Supersymmetry and randomness turn out to be two views of one structure.

Why care? Because flattening a theory to a free Gaussian is the dream outcome. It explains why supersymmetric theories obey unusually strong rules: many quantities that would otherwise receive corrections receive none, and the counting of ground states becomes a robust whole number.

Visual Beginner

A schematic with three panels. The left panel shows a curved bosonic landscape with a noisy trajectory rolling and settling — the Langevin picture. The middle panel shows an arrow labelled "rename" carrying that landscape to the right panel, a flat parabolic bowl, the free Gaussian. A small box beneath the arrow shows two determinants, the rename's stretching factor and the fermion factor, joined by an equals sign and cancelling.

The picture captures the whole idea: the rename flattens the bosonic action to a free Gaussian, and the stretching factor of the rename equals the fermion factor, so the two cancel and only the flat bowl is left.

Worked example Beginner

Take supersymmetric quantum mechanics on a time line and watch the rename in action on a one-dimensional path.

Step 1. The bosonic variable is a path $x(t)$ in time. The theory comes with a superpotential, a function $W(x)$. The bosonic weight depends on the combination "velocity plus the slope of $W$," written $\dot{x}+W^{\prime }(x)$, where $W^{\prime }$ is the derivative of $W$. The fermions contribute a separate factor.

Step 2. Define the new variable by the rename $\xi (t)=\dot{x}(t)+W^{\prime }(x(t))$. This is exactly the noisy-ball update read backwards: the random kick equals the velocity plus the downhill drift set by $W^{\prime }$. The drift of the noisy-ball simulation is the slope of the superpotential.

Step 3. After the rename, the bosonic weight becomes the plain bell curve in $\xi$ — the simplest free Gaussian weight, built from $\xi$ squared. The curved landscape has flattened. That is job one.

Step 4. Any rename carries a stretching factor. Here it measures how a small wiggle in $x$ changes $\xi$, and it works out to a factor built from the second derivative $W^{\prime \prime }(x)$. The fermions, when summed away, produce the very same factor. They cancel exactly. That is job two.

What this tells us: the supersymmetric theory, after the rename, is a free Gaussian and nothing more. The rename is the noisy-ball update; the drift is the slope of the superpotential; and the fermion bookkeeping is the stretching factor of the rename. Supersymmetry has become a statement about noise.

Check your understanding Beginner

Formal definition Intermediate+

Let a Euclidean supersymmetric theory have bosonic fields $\varphi$ and fermionic fields $\psi ,\bar{\psi }$, with action $S[\varphi ,\psi ,\bar{\psi }]=S_B[\varphi ]+\bar{\psi }D[\varphi ]\psi$, where $S_B$ is the bosonic action and $D[\varphi ]$ is the Dirac/Yukawa operator whose entries depend on $\varphi$. Integrating out the fermions by the Berezin/Grassmann rule of 08.14.02 gives the effective bosonic measure

$$d\mu (\varphi )=\det D[\varphi ]e^{-S_B[\varphi ]}\mathcal{D}\varphi .$$

A Nicolai map is a (generally nonlinear, nonlocal) change of bosonic variables $\xi =\xi [\varphi ]$, with local inverse $\varphi =\varphi [\xi ]$, such that two conditions hold simultaneously:

$$\text{(i)}{S}_B[\varphi ]⟼\frac{1}{2}\int \xi [\varphi {]}^2,\text{(ii)}\det \left(\frac{\delta \xi [\varphi ]}{\delta \varphi }\right)=\det D[\varphi ].$$

Condition (i) says the bosonic action becomes free and Gaussian in $\xi$; condition (ii) says the functional Jacobian of the map equals the fermion determinant. Together they give

$$d\mu (\varphi )=\det D[\varphi ]e^{-S_B[\varphi ]}\mathcal{D}\varphi =\det \left(\frac{\delta \xi }{\delta \varphi }\right)e^{-\frac{1}{2}\int {\xi }^2}\mathcal{D}\varphi =e^{-\frac{1}{2}\int {\xi }^2}\mathcal{D}\xi ,$$

so the partition function $Z=\int d\mu (\varphi )$ collapses to a free Gaussian integral. The existence of such a map is the defining characterisation of supersymmetry due to Nicolai 1980.

The stochastic-quantisation reading identifies $\xi$ with the Gaussian noise. For a theory with bosonic action $S_B[\varphi ]=\frac{1}{2}\int (\dot{\varphi }+W^{\prime }(\varphi ){)}^2$ (the form forced by supersymmetry, with $W$ the superpotential), the Langevin/drift relation

$$\xi =\dot{\varphi }+W^{\prime }(\varphi )⟺\dot{\varphi }=-W^{\prime }(\varphi )+\xi$$

is precisely the Nicolai map: $\xi$ is the white noise of the Langevin equation, the drift is the superpotential gradient $-W^{\prime }$, and condition (i) holds by construction since $S_B=\frac{1}{2}\int {\xi }^2$. The Jacobian in condition (ii) is then the functional determinant of the linearised drift operator,

$$\frac{\delta \xi (t)}{\delta \varphi (t^{\prime })}=\left(\frac{d}{dt}+W^{\prime \prime }(\varphi (t))\right)\delta (t-t^{\prime }),\det \left(\frac{d}{dt}+W^{\prime \prime }(\varphi )\right)=\det D[\varphi ],$$

which equals the fermion determinant from integrating out $\psi ,\bar{\psi }$.

Counterexamples to common slips

• Not every change of variables is a Nicolai map. Condition (i) alone (flattening the bosonic action) can be arranged for many theories; what makes the map special is that the same map also satisfies (ii), and (ii) holds only because the theory is supersymmetric. A non-supersymmetric theory generically has $\det (\delta \xi /\delta \varphi )\ne \det D[\varphi ]$, and the cancellation fails.
• The Jacobian sign and phase matter. The determinant $\det ({\partial }_t+W^{\prime \prime }(\varphi ))$ requires a regularisation (zeta-function or heat-kernel); its regularised phase / winding is where the Witten index $\mathrm{tr}(-1{)}^F$ of 08.10.11 enters. Dropping the regularisation discards the topological information that distinguishes broken from unbroken supersymmetry.
• The map is nonlocal in general. In field theory $\xi [\varphi ]$ is a nonlocal functional of $\varphi$; only in quantum mechanics (one time dimension) is the drift relation the simple local first-order equation $\xi =\dot{\varphi }+W^{\prime }(\varphi )$.

Key theorem with proof Intermediate+

Theorem (Nicolai map for supersymmetric quantum mechanics; Nicolai 1980 ^{[Nicolai-PLB]}, Cecotti-Girardello 1983 ^{[Cecotti-Girardello]}). Let supersymmetric quantum mechanics on the Euclidean time line have bosonic action $S_B[x]=\frac{1}{2}\int dt(\dot{x}+W^{\prime }(x){)}^2$ and fermion bilinear $\bar{\psi }({\partial }_t+W^{\prime \prime }(x))\psi$, with $W$ the superpotential. Define the change of variables $\xi (t)=\dot{x}(t)+W^{\prime }(x(t))$. Then (i) $S_B[x]=\frac{1}{2}\int dt\xi (t{)}^2$, and (ii) the Jacobian of the map equals the fermion determinant:

$$\det \left(\frac{\delta \xi }{\delta x}\right)=\det \left({\partial }_t+W^{\prime \prime }(x)\right)=\int \mathcal{D}\bar{\psi }\mathcal{D}\psi e^{-\int dt\bar{\psi }({\partial }_t+W^{\prime \prime }(x))\psi }.$$

Consequently the full partition function reduces to a free Gaussian integral over $\xi$, and the Nicolai variable $\xi$ is the Gaussian white noise of the Langevin equation $\dot{x}=-W^{\prime }(x)+\xi$.

Proof. For (i), substitute the definition: $S_B[x]=\frac{1}{2}\int dt(\dot{x}+W^{\prime }(x){)}^2=\frac{1}{2}\int dt{\xi }^2$ by the rename. This is immediate from the form of the bosonic action, which supersymmetry forces to be a perfect square in $\dot{x}+W^{\prime }(x)$.

For (ii), compute the functional derivative of $\xi (t)=\dot{x}(t)+W^{\prime }(x(t))$ with respect to $x(t^{\prime })$:

$$\frac{\delta \xi (t)}{\delta x(t^{\prime })}=\frac{d}{dt}\delta (t-t^{\prime })+W^{\prime \prime }(x(t))\delta (t-t^{\prime })=\left({\partial }_t+W^{\prime \prime }(x(t))\right)\delta (t-t^{\prime }).$$

So the Jacobian operator is the first-order operator ${\partial }_t+W^{\prime \prime }(x)$. The Berezin integration rule of 08.14.02, $\int \mathcal{D}\bar{\psi }\mathcal{D}\psi e^{-\bar{\psi }M\psi }=\det M$, applied to $M={\partial }_t+W^{\prime \prime }(x)$, gives the fermion determinant as exactly this operator determinant. Hence $\det (\delta \xi /\delta x)=\det ({\partial }_t+W^{\prime \prime }(x))=\det D[x]$.

Combining, the change of variables $x\mapsto \xi$ in the measure gives

$$Z=\int \mathcal{D}x\det \left({\partial }_t+W^{\prime \prime }(x)\right)e^{-S_B[x]}=\int \mathcal{D}x\det \left(\frac{\delta \xi }{\delta x}\right)e^{-\frac{1}{2}\int {\xi }^2}=\int \mathcal{D}\xi e^{-\frac{1}{2}\int {\xi }^2},$$

a free Gaussian integral. The rearrangement $\xi =\dot{x}+W^{\prime }(x)\iff \dot{x}=-W^{\prime }(x)+\xi$ identifies $\xi$ as the white noise of the Langevin equation with drift $-W^{\prime }(x)$, the superpotential gradient. $\square$

Bridge. This theorem builds toward the whole supersymmetric-Langevin dictionary and appears again in the Master-tier treatment of field-theoretic Nicolai maps and the Parisi-Sourlas noise-as-superpartner construction. The central insight is exactly that the fermion determinant and the change-of-variable Jacobian are the same object: this is exactly the foundational reason supersymmetric theories are characterised by the existence of a noise change of variables, and putting these together with the stochastic-quantisation framework of 08.10.08 shows that the SUSY-QM superpotential $W$ IS the Langevin drift. The map generalises from one time dimension, where the drift relation is the local equation $\xi =\dot{x}+W^{\prime }(x)$, to field theory, where $\xi [\varphi ]$ is a nonlocal functional but the Jacobian-equals-determinant identity persists. The bridge is the observation that the regularised phase of $\det ({\partial }_t+W^{\prime \prime }(x))$ carries the Witten index of 08.10.11, so the topological counting of supersymmetric ground states and the winding of the Langevin drift map are one and the same datum.

Exercises Intermediate+

Advanced results Master

Theorem (existence of the Nicolai map characterises supersymmetry; Nicolai 1980 ^{[Nicolai-NPB]}, Cecotti-Girardello 1983 ^{[Cecotti-Girardello]}). A scalar theory with bosonic action $S_B[\varphi ]$ and fermion content integrating to $\det D[\varphi ]$ is supersymmetric if and only if there exists a change of bosonic variables $\xi =\xi [\varphi ]$ satisfying both $S_B[\varphi ]=\frac{1}{2}\int \xi [\varphi {]}^2$ and $\det (\delta \xi /\delta \varphi )=\det D[\varphi ]$. The map renders the partition function a free Gaussian integral and exists order by order in perturbation theory.

The map is constructed perturbatively. Writing $\xi [\varphi ]={\xi }_0[\varphi ]+g{\xi }_1[\varphi ]+g^2{\xi }_2[\varphi ]+\cdots$ in the coupling $g$, the conditions (i) and (ii) at each order determine ${\xi }_n$ from the lower orders; Dietz-Lechtenfeld 1985 ^{[Dietz-Lechtenfeld]} carry this out and show the Jacobian terms reproduce, order by order, the fermion-loop diagrams that integrating out $\psi ,\bar{\psi }$ would produce. The free Gaussian form means correlators of $\xi$ are products of free propagators (Wick's theorem), and the entire content of any boson correlator $⟨\varphi \cdots \varphi ⟩$ comes entirely from the inverse map $\varphi [\xi ]$.

Theorem (Parisi-Sourlas: noise as a supersymmetric partner; Parisi-Sourlas 1982 ^{[Parisi-Sourlas]}). The Onsager-Machlup weight of a Langevin process $\dot{\varphi }=-W^{\prime }(\varphi )+\xi$, after integrating out the Gaussian noise $\xi$, equals the action of a supersymmetric theory in which the noise is the bosonic superpartner of a fermion pair $\psi ,\bar{\psi }$ supplying the Jacobian determinant. The hidden supersymmetry has supercharges that rotate the bosonic noise field into the fermions and back.

The construction: the Langevin equation defines a delta-functional $\delta (\dot{\varphi }+W^{\prime }(\varphi )-\xi )$ enforcing the equation of motion; representing the delta-functional by an auxiliary field and exponentiating the Jacobian $\det ({\partial }_t+W^{\prime \prime }(\varphi ))$ by Grassmann fields $\psi ,\bar{\psi }$ produces an action invariant under a supersymmetry that pairs the noise with the fermions. This is the field-theoretic content of the Nicolai map seen from the stochastic side: the existence of a clean noise change of variables is the same statement as the existence of the hidden supersymmetry. Parisi and Sourlas used the same mechanism to derive dimensional reduction for random-field systems, the subject of 08.10.13.

Theorem (the Witten index as the regularised Jacobian phase; Witten 1982 ^[Witten], Cecotti-Girardello 1983 ^{[Cecotti-Girardello]}). The functional determinant $\det ({\partial }_t+W^{\prime \prime }(x))$ of the Nicolai-map Jacobian requires regularisation; its regularised value carries a phase / sign equal to $(-1)$ raised to the spectral-flow / winding number of the drift map, and this datum is the Witten index $\mathrm{tr}(-1{)}^F$ of 08.10.11. Supersymmetry is unbroken if and only if the index is nonzero, which corresponds to the drift $W^{\prime }(x)$ changing sign at infinity (odd-degree superpotential behaviour).

The index counts $n_B^{E=0}-n_F^{E=0}$, the difference of bosonic and fermionic zero-energy states, and is invariant under continuous deformations of $W$ that preserve the behaviour at infinity. In the Nicolai-map language it is the winding number of $W^{\prime }(x)$ as $x$ runs over the real line — the number of times the drift crosses zero with a definite orientation — which is exactly the regularised phase of $\det ({\partial }_t+W^{\prime \prime }(x))$. The map thus encodes the topological ground-state counting in the analytic structure of the Jacobian.

Theorem (lattice route to supersymmetry via the Nicolai map; Damgaard-Hüffel 1987 ^{[Damgaard-Huffel]}). Because the Nicolai map replaces the supersymmetric measure by a free Gaussian measure in $\xi$, a lattice discretisation that preserves the map preserves supersymmetry exactly at finite lattice spacing, sidestepping the generic breaking of supersymmetry by the lattice. The constraint is that the discretised drift relation $\xi =\dot{\varphi }+W^{\prime }(\varphi )$ retain an exact discrete Jacobian-determinant identity.

This is the practical payoff: ordinary lattice regularisation breaks supersymmetry because the lattice has no exact translation invariance in superspace, but a Nicolai-map-based discretisation keeps the noise change of variables exact, and with it the cancellation of the bosonic and fermionic contributions. The approach connects to the stochastic-quantisation lattice numerics of 08.10.08: the supersymmetric Langevin process is simulated with the superpotential as the drift, and the fermion determinant is automatically supplied by the Jacobian rather than computed separately.

Synthesis. The Nicolai map putting together supersymmetry and stochastic quantisation is the central insight that a supersymmetric theory is exactly a Langevin process with a clean noise change of variables: the foundational reason is that the Jacobian $\det (\delta \xi /\delta \varphi )$ of the noise map equals the fermion determinant $\det D[\varphi ]$, so integrating out the fermions and changing bosonic variables are the same operation, and this is exactly why the partition function collapses to a free Gaussian. The bridge is the drift relation $\xi =\dot{\varphi }+W^{\prime }(\varphi )$, which generalises the SUSY-QM dictionary of 08.10.11 — superpotential equals Langevin drift — from the operator side to the measure side, and putting these together with the Fokker-Planck framework of 08.10.02 identifies the static equilibrium as the modulus-squared of the supersymmetric ground state. The central insight extends forward to 08.10.13, where the same noise-as-superpartner mechanism of Parisi-Sourlas yields dimensional reduction and the Langevin noise is dual to the quenched random source, and toward lattice supersymmetry, where the map's finite-spacing exactness keeps supersymmetry unbroken. The synthesis is that supersymmetry, the fermion determinant, the Langevin noise, and the Witten index are four faces of one structure — a change of variables that flattens the bosonic action while its Jacobian supplies the fermions — and this is exactly the structure that makes supersymmetric theories analytically and numerically tractable.

Full proof set Master

Proposition (the Nicolai map collapses the SUSY-QM partition function to a Gaussian). For supersymmetric quantum mechanics with bosonic action $S_B[x]=\frac{1}{2}\int dt(\dot{x}+W^{\prime }(x){)}^2$ and fermion bilinear $\bar{\psi }({\partial }_t+W^{\prime \prime }(x))\psi$, the partition function $Z=\int \mathcal{D}x\mathcal{D}\bar{\psi }\mathcal{D}\psi e^{-S}$ equals the free Gaussian integral $\int \mathcal{D}\xi e^{-\frac{1}{2}\int {\xi }^2}$.

Proof. Integrate out the fermions first using the Berezin rule of 08.14.02:

$$\int \mathcal{D}\bar{\psi }\mathcal{D}\psi e^{-\int dt\bar{\psi }({\partial }_t+W^{\prime \prime }(x))\psi }=\det ({\partial }_t+W^{\prime \prime }(x)).$$

The remaining bosonic integral is $Z=\int \mathcal{D}x\det ({\partial }_t+W^{\prime \prime }(x))e^{-\frac{1}{2}\int (\dot{x}+W^{\prime }(x){)}^2}$. Change variables to $\xi (t)=\dot{x}(t)+W^{\prime }(x(t))$. The functional Jacobian of this change of variables is

$$\left|\det \frac{\delta \xi }{\delta x}\right|=\left|\det ({\partial }_t+W^{\prime \prime }(x))\right|,$$

so $\mathcal{D}\xi =\det ({\partial }_t+W^{\prime \prime }(x))\mathcal{D}x$ (the determinant is positive for the regularisation that preserves the index in the unbroken case). Substituting,

$$Z=\int \mathcal{D}x\det ({\partial }_t+W^{\prime \prime }(x))e^{-\frac{1}{2}\int {\xi }^2}=\int \mathcal{D}\xi e^{-\frac{1}{2}\int {\xi }^2}.$$

The fermion determinant has been absorbed by the Jacobian, and the bosonic action is now the free Gaussian weight in $\xi$. $\square$

Proposition (the Nicolai variable is the Langevin white noise). The change of variables $\xi =\dot{x}+W^{\prime }(x)$ identifies $\xi$ as the Gaussian white noise of the Langevin equation $\dot{x}=-W^{\prime }(x)+\xi$, with $⟨\xi (t)\xi (t^{\prime })⟩=\delta (t-t^{\prime })$ induced by the free Gaussian weight $e^{-\frac{1}{2}\int {\xi }^2}$.

Proof. Rearranging $\xi =\dot{x}+W^{\prime }(x)$ gives $\dot{x}=-W^{\prime }(x)+\xi$, the Langevin equation with drift $-W^{\prime }(x)$ and forcing $\xi$. Under the measure $e^{-\frac{1}{2}\int {\xi }^2}\mathcal{D}\xi$, the field $\xi$ is a centred Gaussian with covariance read off from the quadratic form $\frac{1}{2}\int {\xi }^2=\frac{1}{2}\int dt\xi (t{)}^2$, namely $⟨\xi (t)\xi (t^{\prime })⟩=\delta (t-t^{\prime })$. So $\xi$ is white noise, and the Langevin process generated by $\dot{x}=-W^{\prime }(x)+\xi$ has $x$-correlators that, by the Parisi-Wu equilibrium identification of 08.10.08, reproduce the supersymmetric-quantum-mechanics path integral. The drift $-W^{\prime }(x)$ is the gradient of the superpotential. $\square$

Proposition (Witten index as the regularised determinant sign for the harmonic case). For $W(x)=\frac{1}{2}\omega x^2$ with $\omega >0$, the regularised Jacobian determinant $\det ({\partial }_t+\omega )$ has sign giving Witten index $\mathrm{tr}(-1{)}^F=1$ (unbroken supersymmetry), consistent with the single normalisable zero-energy ground state $e^{-\frac{1}{2}\omega x^2}$.

Proof. The operator ${\partial }_t+W^{\prime \prime }(x)={\partial }_t+\omega$ on the time line has, with appropriate boundary conditions, a spectrum whose zeta-regularised determinant is positive, and the drift $W^{\prime }(x)=\omega x$ changes sign once (at $x=0$) with positive slope. The Witten index equals the winding / spectral-flow number of this drift, which is $+1$. On the Hamiltonian side, the partner Hamiltonian $H_{-}=-\frac{1}{2}{\partial }_x^2+\frac{1}{2}{\omega }^2{x}^2-\frac{1}{2}\omega$ has ground state ${\psi }_0=e^{-\frac{1}{2}\omega x^2}$ at energy zero, normalisable, while $H_{+}$ has no zero-energy state; so $n_B^{E=0}-n_F^{E=0}=1-0=1$. The two computations agree: $\mathrm{tr}(-1{)}^F=1$, the regularised phase of the Jacobian. $\square$

Proposition (failure of cancellation without supersymmetry). If the fermion operator is $D={\partial }_t+V(x)$ with $V\ne W^{\prime \prime }$ for the bosonic superpotential $W$, the change of variables $\xi =\dot{x}+W^{\prime }(x)$ has Jacobian $\det ({\partial }_t+W^{\prime \prime }(x))\ne \det D$, so the partition function does not reduce to a free Gaussian.

Proof. The change of variables $\xi =\dot{x}+W^{\prime }(x)$ produces the Jacobian $\det ({\partial }_t+W^{\prime \prime }(x))$ regardless of the fermion content, as computed in the key theorem. The fermion integration produces $\det D=\det ({\partial }_t+V(x))$. These determinants are equal as functionals of $x$ only if $W^{\prime \prime }(x)=V(x)$ for all $x$, the supersymmetry condition. When $V\ne W^{\prime \prime }$, the ratio $\det ({\partial }_t+W^{\prime \prime }(x))/\det ({\partial }_t+V(x))$ is a residual functional of $x$, leaving $Z=\int \mathcal{D}\xi e^{-\frac{1}{2}\int {\xi }^2}[\det ({\partial }_t+V)/\det ({\partial }_t+W^{\prime \prime })]$, which is not the free Gaussian. The cancellation is exactly the supersymmetry condition $V=W^{\prime \prime }$. $\square$

Connections Master

• Supersymmetric quantum mechanics 08.10.11. The superpotential $W(x)$ of Hamiltonian SUSY-QM is the Langevin drift potential of the Nicolai map: the partner Hamiltonians $H_{\pm }=-\frac{1}{2}{\partial }_x^2+\frac{1}{2}{W}^{\prime 2}\mp \frac{1}{2}{W}^{\prime \prime }$ have $W^{\prime \prime }$ as the fermion-operator entry, which is exactly the linearised drift ${\partial }_t+W^{\prime \prime }(x)$ whose determinant the Nicolai Jacobian supplies. The Witten index $\mathrm{tr}(-1{)}^F$ of 08.10.11 reappears here as the regularised phase / winding of that Jacobian determinant, and the SUSY ground state $e^{-W}$ is the square root of the Langevin equilibrium density.

• Fokker-Planck equation and equilibrium distribution 08.10.02. The static face of the Nicolai map is the Fokker-Planck equilibrium: the Langevin equation $\dot{x}=-W^{\prime }(x)+\xi$ has stationary density $e^{-2W(x)}/Z$, which equals the modulus-squared $|{\psi }_0{|}^2$ of the supersymmetric ground state. The drift-equals-superpotential-gradient identification is the bridge from the operator-side SUSY-QM to the measure-side stochastic process, and the spectral-gap machinery of 08.10.02 controls the approach to equilibrium.

• Langevin updates and lattice numerics 08.10.08. The Nicolai variable $\xi$ is precisely the Gaussian white noise of the Parisi-Wu Langevin equation of 08.10.08, with the superpotential as the drift. This makes the Nicolai map the supersymmetric instance of stochastic quantisation, and it opens a lattice-friendly route to supersymmetric theories: a discretisation preserving the noise change of variables keeps the bosonic-fermionic cancellation exact at finite spacing.

• Grassmann integration and the 2D Ising model as free fermions 08.14.02. The fermion determinant that the Nicolai Jacobian cancels is computed by the Berezin/Grassmann rule $\int d\bar{\psi }d\psi e^{-\bar{\psi }M\psi }=\det M$ of 08.14.02, applied to the Yukawa/Dirac operator $M={\partial }_t+W^{\prime \prime }(x)$. The whole Nicolai construction rests on this identity: the fermionic integration produces a determinant, and the bosonic change of variables produces the matching Jacobian.

• Parisi-Sourlas dimensional reduction 08.10.13. The noise-as-superpartner mechanism that underlies the Nicolai map is the same mechanism Parisi and Sourlas use to derive dimensional reduction of random-field systems: the Langevin noise is dual to a quenched random source, and the hidden supersymmetry of the stochastic process becomes the $\mathrm{OSp}$ superrotation of the random-field problem. This unit supplies the noise-change-of-variables half of that story; 08.10.13 supplies the random-source half.

Historical & philosophical context Master

Hermann Nicolai introduced the map in two 1980 papers, On a new characterization of scalar supersymmetric theories (Phys. Lett. B 89, 341) ^{[Nicolai-PLB]} and Supersymmetry and functional integration measures (Nucl. Phys. B 176, 419) ^{[Nicolai-NPB]}. His motivation was structural: he sought an intrinsic characterisation of supersymmetry at the level of the bosonic functional measure, without reference to the explicit fermionic fields. The answer — that supersymmetric theories are exactly those admitting a bosonic change of variables flattening the action to a free Gaussian whose Jacobian equals the fermion determinant — recast supersymmetry as a statement about measures rather than about symmetry transformations. Sergio Cecotti and Luciano Girardello, in Functional measure, topology, and dynamical supersymmetry breaking (Ann. Phys. 145, 81, 1983) ^{[Cecotti-Girardello]}, developed the existence and structure of the map, the role of boundary conditions, and the appearance of the Witten index of Edward Witten's 1982 Constraints on supersymmetry breaking (Nucl. Phys. B 202, 253) ^[Witten] as the regularised topological phase of the Jacobian determinant.

The stochastic-quantisation face of the map emerged in parallel. Giorgio Parisi and Nicolas Sourlas, in Supersymmetric field theories and stochastic differential equations (Nucl. Phys. B 206, 321, 1982) ^{[Parisi-Sourlas]}, showed that the noise of a Langevin equation is the bosonic partner of an emergent fermion pair, so that every stochastic differential equation hides a supersymmetry — the same supersymmetry the Nicolai map makes explicit. The identification of the Nicolai variable with the Langevin noise, and of the superpotential with the Langevin drift, ties the supersymmetric structure of Nicolai 1980 to the stochastic quantisation of Parisi and Wu 1981. Klaus Dietz and Olaf Lechtenfeld, in Nicolai maps and stochastic observables (Nucl. Phys. B 259, 397, 1985) ^{[Dietz-Lechtenfeld]}, constructed the map order by order in the coupling and exhibited the fermion-loop cancellations as Jacobian terms. The canonical synthesis is the §6 of Poul Damgaard and Helmuth Hüffel's 1987 review Stochastic quantization (Phys. Rep. 152, 227) ^{[Damgaard-Huffel]}, which reads the map as the equilibrium change of variables of the supersymmetric Langevin process. Philosophically, the map is striking because it makes supersymmetry — a symmetry usually presented through anticommuting transformations — synonymous with the existence of a clean noise change of variables in a stochastic process: supersymmetry and randomness, two ideas from distant corners of physics, turn out to be one structure viewed from two sides.

Bibliography Master

@article{Nicolai1980PLB,
  author  = {Nicolai, Hermann},
  title   = {On a new characterization of scalar supersymmetric theories},
  journal = {Physics Letters B},
  volume  = {89},
  year    = {1980},
  pages   = {341--346}
}

@article{Nicolai1980NPB,
  author  = {Nicolai, Hermann},
  title   = {Supersymmetry and functional integration measures},
  journal = {Nuclear Physics B},
  volume  = {176},
  year    = {1980},
  pages   = {419--428}
}

@article{CecottiGirardello1983,
  author  = {Cecotti, Sergio and Girardello, Luciano},
  title   = {Functional measure, topology, and dynamical supersymmetry breaking},
  journal = {Annals of Physics},
  volume  = {145},
  year    = {1983},
  pages   = {81--99}
}

@article{ParisiSourlas1982,
  author  = {Parisi, Giorgio and Sourlas, Nicolas},
  title   = {Supersymmetric field theories and stochastic differential equations},
  journal = {Nuclear Physics B},
  volume  = {206},
  year    = {1982},
  pages   = {321--332}
}

@article{DietzLechtenfeld1985,
  author  = {Dietz, Klaus and Lechtenfeld, Olaf},
  title   = {Nicolai maps and stochastic observables},
  journal = {Nuclear Physics B},
  volume  = {259},
  year    = {1985},
  pages   = {397--412}
}

@article{Witten1982,
  author  = {Witten, Edward},
  title   = {Constraints on supersymmetry breaking},
  journal = {Nuclear Physics B},
  volume  = {202},
  year    = {1982},
  pages   = {253--316}
}

@article{DamgaardHuffel1987,
  author  = {Damgaard, Poul H. and H{\"u}ffel, Helmuth},
  title   = {Stochastic quantization},
  journal = {Physics Reports},
  volume  = {152},
  number  = {5-6},
  year    = {1987},
  pages   = {227--398}
}

@article{CooperKhareSukhatme1995,
  author  = {Cooper, Fred and Khare, Avinash and Sukhatme, Uday},
  title   = {Supersymmetry and quantum mechanics},
  journal = {Physics Reports},
  volume  = {251},
  year    = {1995},
  pages   = {267--385}
}

@article{ParisiWu1981,
  author  = {Parisi, Giorgio and Wu, Yong-Shi},
  title   = {Perturbation theory without gauge fixing},
  journal = {Scientia Sinica},
  volume  = {24},
  year    = {1981},
  pages   = {483--496}
}

@book{Junker1996,
  author    = {Junker, Georg},
  title     = {Supersymmetric Methods in Quantum and Statistical Physics},
  publisher = {Springer-Verlag},
  year      = {1996}
}