Matchgate Commutant in Free-Fermion Systems

• Matchgate commutants are algebras of operators invariant under replicated free-fermionic (Gaussian) unitaries, defined via bridge operators and Majorana pairing.
• They leverage explicit symmetry structures and replica invariance to yield polynomially growing dimensions and representations of the so(k) Lie algebra.
• These invariants simplify doped matchgate circuit dynamics and inform state frame potentials, providing dual algebraic and geometric formulations.

Searching arXiv for papers on matchgate commutants and related free-fermion commutant structure. The matchgate commutant is the algebra of operators invariant under replicated actions of matchgate, or fermionic Gaussian, unitaries. In one line of work, this notion is formulated for $k$ replicas as

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$

where $M_n$ is the matchgate group on $n$ qubits (Trigueros et al., 22 Jun 2026). In another, related usage, the topic appears as a permutation-stability problem for matchgate signatures: for a fixed signature $f$, one studies the subgroup of variable permutations that preserve matchgateness, and the class of signatures whose entire $S_n$-orbit remains inside the matchgate variety $\mathscr M$ (Meng et al., 27 Mar 2025). These two notions are distinct but structurally parallel: both isolate the symmetries that remain compatible with the Pfaffian or free-fermionic constraints defining matchgates.

1. Matchgates, Majoranas, and replica invariance

Matchgate circuits are the free-fermion, or fermionic Gaussian, sector of qubit dynamics. In the Majorana description, a chain of $n$ qubits is mapped to $N=2n$ Majorana modes $\{\gamma_\mu\}$ via the Jordan–Wigner transform, with $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$0, and the matchgate group $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$1 is generated by bilinears $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$2 (Trigueros et al., 22 Jun 2026). Equivalently, a unitary $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$3 acts linearly on Majoranas,

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$4

so matchgate dynamics is represented by orthogonal transformations on the Majorana space (Trigueros et al., 22 Jun 2026).

This orthogonal action is the algebraic core of the commutant. In the superoperator picture based on Majorana monomials $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$5, matchgate conjugation preserves the Majorana degree $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$6, and the superoperator decomposes as a direct sum of compound matrices $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$7 indexed by degree sectors $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$8 (Burkat et al., 2024). This already implies a minimal commutant generated by the degree projectors $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$9, since these projectors commute with every matchgate superoperator (Burkat et al., 2024).

The replicated commutant formalizes the same symmetry at higher moment order. For the representation $M_n$0, the matchgate twirl

$M_n$1

is an orthogonal projector whose image is exactly $M_n$2 (Trigueros et al., 22 Jun 2026). The problem is therefore to determine the invariant algebra of $M_n$3-copy operators under all Gaussian orthogonal actions.

A closely related single-copy perspective appears in the study of matchgate signatures. There, the relevant symmetry is not replica invariance but invariance under variable permutations. For a matchgate signature $M_n$4 of arity $M_n$5, one defines

$M_n$6

and calls $M_n$7 permutable when this stabilizer is the full symmetric group $M_n$8 (Meng et al., 27 Mar 2025). This suggests a finite-dimensional “commutant under permutations” viewpoint for signatures, complementary to the replica commutant for Gaussian unitaries.

2. Two-copy commutant and the bridge operator

For $M_n$9, the matchgate commutant is completely explicit. On two replicas one defines the bridge operator

$n$0

with $n$1 (Trigueros et al., 22 Jun 2026). Its eigenvalues are

$n$2

and the eigenspace for $n$3 has dimension $n$4 (Trigueros et al., 22 Jun 2026). Let $n$5 denote the corresponding spectral projectors. Then the two-copy matchgate twirl takes the diagonal form

$n$6

so the commutant is exactly the span of the $n$7, and

$n$8

(Trigueros et al., 22 Jun 2026).

The same two-copy space also admits a basis adapted to Majorana weights: $n$9 which yields another basis of $f$0 (Trigueros et al., 22 Jun 2026). In this formulation, each invariant is a uniform pairing over identical Majorana monomials of fixed weight.

A distinct but compatible formulation appears in the general theory of the matchgate commutant. There the basic invariant operators are built from pairings of replicated Majorana slots, and for $f$1 the orthonormal basis is

$f$2

again with $f$3 (Sierant et al., 12 Mar 2026). This identifies the two-copy commutant simultaneously as the bridge spectral algebra, a Majorana-weight pairing algebra, and the degree-projector algebra seen in superoperator form.

The two-copy commutant also plays a central operational role. In doped matchgate circuits, each Gaussian layer acts as $f$4, projecting dynamics back to the commutant, and the remaining evolution is governed by a finite-dimensional transfer matrix in the projector basis $f$5 (Trigueros et al., 22 Jun 2026). Thus, for $f$6, the commutant is not merely an invariant algebra but the exact reduced state space of moment dynamics.

3. General $f$7: $f$8, bridge algebras, and dimension formulas

For arbitrary replica number $f$9, the decisive structural result is that operators coupling different replicas generate an orthogonal Lie algebra in replica space. In the theory of the matchgate commutant, one introduces bridge operators

$S_n$0

which satisfy

$S_n$1

so they furnish a representation of $S_n$2 (Sierant et al., 12 Mar 2026). The central theorem is

$S_n$3

namely that the matchgate commutant is exactly the image of the universal enveloping algebra generated by these bridges (Sierant et al., 12 Mar 2026).

This bridge-algebra description resolves the higher-$S_n$4 invariant problem. The commutant decomposes into irreducible $S_n$5 sectors,

$S_n$6

with admissible highest weights $S_n$7 determined by $S_n$8 and $S_n$9 (Sierant et al., 12 Mar 2026). The resulting dimension formula is

$\mathscr M$0

which grows polynomially in $\mathscr M$1 for fixed $\mathscr M$2 (Sierant et al., 12 Mar 2026). This polynomial growth is one of the main reasons the commutant becomes a practical computational tool.

A complementary geometric description reaches the same object from a different angle. For Majorana free fermions, the $\mathscr M$3-commutant is identified with the ground-state space of a ferromagnetic Heisenberg model with $\mathscr M$4 symmetry in replica space (Lastres et al., 6 Apr 2026). On each physical site the local replica degrees of freedom form spinor irreducible representations of $\mathscr M$5, and the global ground space splits into two irreducible sectors $\mathscr M$6 and $\mathscr M$7, exchanged by parity, so that the full commutant is a single irreducible $\mathscr M$8 object when parity is included (Lastres et al., 6 Apr 2026). This does not contradict the $\mathscr M$9 bridge description: it suggests different organizing symmetries for the same invariant space, with $n$0 arising from pairwise replica couplings and $n$1 arising from the ferromagnetic replica-space geometry.

The geometric framework also gives a closed dimension formula in terms of Weyl dimensions for the replica-space ferromagnetic irreducible sectors (Lastres et al., 6 Apr 2026). This suggests that the algebraic basis construction of (Sierant et al., 12 Mar 2026) and the coherent-state geometry of (Lastres et al., 6 Apr 2026) are two compatible resolutions of the same replica-invariant structure.

4. Explicit bases, Gelfand–Tsetlin resolution, and geometry

At low replica number, basis elements can be written directly. For $n$2, an orthonormal basis is indexed by $n$3 and has the form

$n$4

where the $n$5 are disjoint subsets of $n$6 of prescribed sizes (Sierant et al., 12 Mar 2026). For $n$7, pairing operators become nonorthogonal and overcomplete, and a refined construction is required.

The general solution uses a Gelfand–Tsetlin chain

$n$8

together with Casimir operators at each level (Sierant et al., 12 Mar 2026). Joint eigenspaces of the commuting family of Casimirs are one-dimensional and indexed by Gelfand–Tsetlin patterns. For each irreducible sector $n$9, diagonal projectors $N=2n$0 onto these one-dimensional sectors are polynomials in the Casimirs, and off-diagonal matrix units are obtained by sandwiching suitable products of ladder operators between such projectors,

$N=2n$1

The resulting operators form an explicit orthonormal basis of $N=2n$2 for all $N=2n$3 and $N=2n$4 (Sierant et al., 12 Mar 2026).

The geometric picture replaces this orthonormal-basis viewpoint by a coherent-state manifold. In the Majorana case, the single-site condition for the commutant ground-state manifold is

$N=2n$5

which is Bravyi’s characterization of fermionic Gaussian states (Lastres et al., 6 Apr 2026). Consequently, the manifold of coherent states underlying the commutant is the orthogonal Grassmannian

$N=2n$6

with even and odd parity components corresponding to $N=2n$7 orbits (Lastres et al., 6 Apr 2026). This yields a compact projection formula

$N=2n$8

where $N=2n$9 projects onto the commutant and $\{\gamma_\mu\}$0 (Lastres et al., 6 Apr 2026). A plausible implication is that the Gelfand–Tsetlin basis and the Grassmannian coherent-state resolution are dual computational realizations of the same projector.

In the superoperator setting, the degree projectors $\{\gamma_\mu\}$1 already provide a coarse commutant algebra generated by block scalars on invariant subspaces $\{\gamma_\mu\}$2 (Burkat et al., 2024). The later works sharpen this into a full basis and full geometry, respectively.

5. Clifford–matchgate subgroup and divergence from the Gaussian commutant

The Clifford–matchgate group

$\{\gamma_\mu\}$3

acts on Majoranas as signed permutations,

$\{\gamma_\mu\}$4

so its symmetry is discrete rather than continuous (Sierant et al., 12 Mar 2026). The corresponding commutant is therefore different.

For $\{\gamma_\mu\}$5 replicas, an invariant is specified by replica patterns $\{\gamma_\mu\}$6 recording on which replicas a given Majorana mode appears, with only even-cardinality subsets surviving due to sign-flip symmetry (Sierant et al., 12 Mar 2026). Let

$\{\gamma_\mu\}$7

and let $\{\gamma_\mu\}$8 count the number of modes realizing pattern $\{\gamma_\mu\}$9, subject to

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$00

The orthogonal basis elements are operators $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$01 built by summing over all signed-permutation orbits with fixed counts $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$02 (Sierant et al., 12 Mar 2026). Their number is

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$03

which scales as $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$04 for fixed $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$05 (Sierant et al., 12 Mar 2026).

The discrete and continuous commutants coincide only for $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$06. For $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$07, new even-cardinality replica patterns of size $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$08 appear in the Clifford–matchgate case, with no continuous Gaussian analogue, and the dimensions diverge (Sierant et al., 12 Mar 2026). This qualitative divergence is also visible in design behavior: Clifford–matchgate circuits form exact matchgate $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$09-designs for $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$10, but not at $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$11 (Sierant et al., 12 Mar 2026).

This replica-order threshold has a natural interpretation. For low replica number, pairwise contractions exhaust all invariants, so the discrete signed-permutation subgroup and the continuous orthogonal group have the same invariant algebra. At higher replica number, discrete orbit data carries additional information not reducible to pairings.

6. Applications and relation to permutation-based “commutants” of signatures

The commutant becomes operational through twirling. For any group $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$12, the $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$13-twirl

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$14

is the orthogonal projector onto $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$15 (Sierant et al., 12 Mar 2026). In the matchgate case, expansion in the orthonormal basis $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$16 gives an explicit twirling channel for all $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$17 (Sierant et al., 12 Mar 2026). For the Gaussian state orbit, the $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$18-copy moment operator of the vacuum orbit is simply the normalized projector onto the trivial $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$19 sector,

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$20

which functions as a fermionic analogue of Weingarten calculus (Sierant et al., 12 Mar 2026).

The commutant also determines unitary and state frame potentials. In particular, the $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$21-th unitary frame potential of matchgates is simply

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$22

and the state frame potential of the Gaussian orbit is the inverse trace of the trivial-sector projector (Sierant et al., 12 Mar 2026). Closed formulas then lead to consequences such as the failure of Clifford–matchgate ensembles to reproduce fourth moments of the full Gaussian orbit with constant relative error (Sierant et al., 12 Mar 2026).

Further applications include a fermionic de Finetti theorem for Gaussian-symmetric states, average nonstabilizerness of random fermionic Gaussian states, and systematic higher-copy non-Gaussianity measures defined by invariant functionals

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$23

all made explicit by the commutant basis (Sierant et al., 12 Mar 2026). The geometric formulation likewise yields coherent-state integral formulas for averaged Rényi entropies of Gaussian states, described as “free-fermion Page curves” (Lastres et al., 6 Apr 2026).

A different, but terminologically related, use of “matchgate commutant” occurs for Boolean matchgate signatures under variable permutations. There the invariant object is not $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$24 but the stabilizer

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$25

and the maximal case consists of permutable matchgate signatures $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$26 (Meng et al., 27 Mar 2025). For normalized signatures $$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$27, permutability is equivalent to

$$Com_k(M_n)=\{X\in \mathrm{End}(\mathcal H_n^{\otimes k}) : [U^{\otimes k},X]=0\ \forall U\in M_n\},$$28

and such signatures admit a star-gadget realization from symmetric matchgate components (Meng et al., 27 Mar 2025). This suggests a finite permutation-centralizer analogue of the replica commutant: in both settings, matchgate structure survives only for highly constrained invariant tensors.

Taken together, these results establish the matchgate commutant as a precise invariant algebra for free-fermion quantum dynamics and, in the signature setting, as a symmetry constraint on Pfaffian tensor structure. In the replica theory, the decisive conclusion is that matchgate invariants are governed not by symmetric-group permutations as in full Haar theory, but by orthogonal replica symmetries generated by bridge operators, with explicit algebraic bases, Grassmannian geometry, and polynomially growing dimensions (Sierant et al., 12 Mar 2026).