Centered Trace-Dobrushin Coefficient in Quantum Channels

• Centered Trace-Dobrushin Coefficient is a contraction measure for positive trace-preserving maps on trace-zero self-adjoint matrices, quantifying residual input dependence.
• It underpins the analysis of quantum channels by linking trace-norm contraction to asymptotic forgetting and replacement channel behavior in both deterministic and random settings.
• It bridges abstract Dobrushin frameworks with quantum dynamics, providing rigorous criteria for memory loss and boundary state convergence in finite-dimensional matrix algebras.

The centered trace-Dobrushin coefficient is a Dobrushin-type contraction coefficient defined for positive trace-preserving maps on finite-dimensional matrix algebras by restricting the channel to the real trace-zero self-adjoint subspace and measuring its operator norm in trace norm. In the explicit terminology of the literature, this object is defined for products of quantum channels and used to quantify the residual dependence of the output on the input state; its decay is the criterion for trace-norm forgetting and for asymptotic replacement by a moving replacement channel (Pathirana, 30 Apr 2026). Closely related antecedents appear in generalized Dobrushin coefficients relative to a projection on abstract state spaces and in cone-based noncommutative analogues of Dobrushin’s ergodicity coefficient, where the relevant norm becomes the trace norm in the quantum setting (Mukhamedov et al., 2020, Gaubert et al., 2013).

1. Formal definition and equivalent characterizations

For a positive and trace-preserving map $T:M_d(\mathbb C)\to M_d(\mathbb C)$, the centered trace-Dobrushin coefficient is defined on the real trace-zero self-adjoint subspace

$$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$

by

$$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$

Equivalently, it is the operator norm of $T$ restricted to $(H,\|\cdot\|_1)$. For a product of channels $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$, the product coefficient is

$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$

This formulation makes the adjective centered literal: the coefficient ignores multiples of the identity by working on the trace-zero sector, which is precisely the space of deviations from a state-independent output (Pathirana, 30 Apr 2026).

The coefficient admits several exact reformulations. For any positive trace-preserving $T$,

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$

It also satisfies

$$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$

Hence $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$0, and

$$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$1

Thus vanishing of the coefficient is equivalent to $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$2 being a replacement channel. For products,

$$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$3

so $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$4 is exactly the trace-norm diameter of the evolved state set $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$5 (Pathirana, 30 Apr 2026).

These identities place the coefficient simultaneously in operator-theoretic, probabilistic, and distinguishability-based frameworks. As an operator norm, it measures contraction on centered observables or centered states; as a state-space diameter, it measures the worst residual input dependence after evolution.

2. Deterministic channel products and asymptotic replacement

For deterministic products of channels, the centered trace-Dobrushin coefficient is the exact product-level measure of residual memory. The defining asymptotic property is trace-norm forgetting: $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$6 which is equivalent to

$$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$7

Small $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$8 therefore means that all input states are sent to nearly the same output state (Pathirana, 30 Apr 2026).

A replacement channel with center $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$9 is

$$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$0

For a fixed reference state $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$1, the moving reference replacement channel is

$$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$2

The key quantitative estimate is

$$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$3

Accordingly, decay of $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$4 is equivalent, up to universal constants, to convergence of the product channel to a replacement channel (Pathirana, 30 Apr 2026).

The deterministic theory identifies four equivalent formulations for fixed $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$5:

1. $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$6 as $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$7.
2. $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$8 is strongly asymptotically replacing from $$\kappa_{\rm tr}(T):=\sup_{X\in H\setminus\{0\}}\frac{\|T(X)\|_1}{\|X\|_1}.$$9: there exist states $T$0 such that

$T$1

1. For every reference state $T$2,

$T$3

2. The same convergence holds for some reference state $T$4.

The significance of these equivalences is structural rather than merely quantitative. They show that the coefficient is not only a bound on distinguishability contraction; it is the intrinsic criterion for whether a non-homogeneous channel product eventually behaves like a rank-one, state-forgetting evolution.

3. Pullback forgetting and boundary states

In the two-sided setting indexed by $T$5, the relevant limit is pullback rather than forward evolution. The pullback memory-loss condition is

$T$6

Under this condition there exists a unique pullback boundary state

$T$7

satisfying

$T$8

Moreover,

$T$9

for every reference state $(H,\|\cdot\|_1)$0, and the limit is independent of $(H,\|\cdot\|_1)$1. The associated canonical replacement family is

$(H,\|\cdot\|_1)$2

Quantitatively,

$(H,\|\cdot\|_1)$3

and

$(H,\|\cdot\|_1)$4

Thus the remote-past limit canonically selects a moving replacement center and upgrades mere forgetting to a boundary-state theory (Pathirana, 30 Apr 2026).

This boundary-state interpretation is one of the main reasons the coefficient is useful beyond one-step contraction analysis. It identifies a dynamically consistent family of effective equilibria even when the channel sequence is time-dependent and non-stationary.

4. Random cocycles, Lyapunov exponents, and convergence rates

For a measurable family of CPTP maps $(H,\|\cdot\|_1)$5 over an invertible probability-preserving system $(H,\|\cdot\|_1)$6, define the cocycle

$(H,\|\cdot\|_1)$7

The basic structural property is submultiplicativity: $(H,\|\cdot\|_1)$8 Equivalently,

$(H,\|\cdot\|_1)$9

By Kingman’s theorem, the trace-Dobrushin Lyapunov exponent

$\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$0

exists almost surely; it is $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$1-invariant, and under ergodicity of $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$2 it is almost surely constant with

$\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$3

The paper proves the equivalence

$\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$4

This identifies negativity of the exponent with quenched trace-memory loss in both forward and pullback directions (Pathirana, 30 Apr 2026).

Assuming $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$5 almost surely, there is a unique dynamically stationary random state

$\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$6

and there exist measurable $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$7-invariant $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$8 and almost surely finite random variables $\Phi_{t:s}:=\Phi_{t-1}\circ\cdots\circ\Phi_s$9 such that

$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$0

$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$1

The state-level estimates have constant $\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$2 instead of $\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$3. The coefficient therefore controls both existence and quantitative rate of replacement mixing (Pathirana, 30 Apr 2026).

A further refinement concerns annealed decay. If the stationary one-step channel process is $\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$4-mixing, in the sense that its maximal-correlation profile $\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$5 tends to zero, then for every $\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$6,

$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$7

If the one-step channels are independent, then

$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$8

This shows that the same coefficient supports both quenched and annealed asymptotic regimes.

5. Relation to earlier Dobrushin frameworks

Before the explicit terminology of centered trace-Dobrushin coefficients for channel products, several closely related Dobrushin constructions had already isolated the same contraction mechanism in different languages. The following comparison captures the main line of descent.

Setting	Quantity	Relation
Product of CPTP maps	$\kappa_{t:s}:=\kappa_{\rm tr}(\Phi_{t:s}) .$9 on $T$0	Explicit named centered trace-Dobrushin coefficient (Pathirana, 30 Apr 2026)
Abstract state spaces	$T$1, $T$2	Projection-centered Dobrushin coefficient; trace-class settings included (Mukhamedov et al., 2020)
Cones and quantum channels	Hopf/Hilbert contraction coefficient with dual trace norm in the matrix case	Noncommutative analogue of Dobrushin’s ergodicity coefficient (Gaubert et al., 2013)
Energy-constrained quantum channels	$T$3	Trace-norm constrained refinement of the ordinary quantum Dobrushin coefficient (Huber et al., 2018)
Measures on countable products	$T$4	Oscillation and coupling formulations coincide exactly (Armstrong-Goodall et al., 2021)

In the abstract-state-space setting, the generalized Dobrushin coefficient relative to a projection $T$5 is

$T$6

When $T$7 is a rank-one Markov projection, this recovers the classical coefficient. In noncommutative $T$8-spaces and trace-class settings, the same theory applies, so $T$9 plays the role of the centered subspace of deviations from equilibrium (Mukhamedov et al., 2020). This suggests that the later centered trace-Dobrushin coefficient is a concrete finite-dimensional quantum realization of a more abstract projection-centered paradigm.

The cone-based formulation sharpens the same point from a dual viewpoint. There the Dobrushin-type quantity is the contraction rate with respect to Hopf’s oscillation seminorm, which is a quotient norm modulo the line $$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$0. In the matrix setting, the dual norm becomes the trace norm, the dual simplex becomes the set of density matrices, and the resulting noncommutative coefficient is explicitly presented as the quantum analogue of Dobrushin’s ergodicity coefficient (Gaubert et al., 2013). The centering mechanism is therefore quotienting out the trivial direction generated by the unit, while the trace aspect enters through the dual norm.

For a single quantum channel, the ordinary trace-norm Dobrushin coefficient is

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$1

The corresponding unconstrained Dobrushin curve is linear: $$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$2 Because this unconstrained curve is trivial, the energy-constrained Dobrushin curve

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$3

becomes the nontrivial refinement (Huber et al., 2018). This is not called centered, but it belongs to the same family of trace-distance contraction quantities.

The measure-theoretic equality

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$4

on countable products of Polish spaces provides another instructive analogue: a Dobrushin metric defined through coordinatewise oscillation/Lipschitz control coincides exactly with a coupling metric defined through optimal joinings (Armstrong-Goodall et al., 2021). A plausible implication is that “oscillation” and “trace-like” formulations are often two representations of the same contraction phenomenon rather than genuinely different invariants.

6. Applications to inhomogeneous matrix product states and terminological scope

The coefficient transfers directly to inhomogeneous matrix product states in left-canonical CPTP gauge, where the auxiliary transfer maps are channels. Writing the right-tail transfer product as

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$5

the condition

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$6

for every fixed $$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$7 implies the existence of a unique right boundary sequence $$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$8 satisfying

$$\kappa_{\rm tr}(T)=\sup_{\substack{\rho,\sigma\in\mathcal S_d\ \rho\neq\sigma}} \frac{\|T(\rho)-T(\sigma)\|_1}{\|\rho-\sigma\|_1} =\frac12\sup_{\rho,\sigma\in\mathcal S_d}\|T(\rho)-T(\sigma)\|_1 .$$9

and a unique infinite-volume state $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$0 on the quasi-local algebra given locally by

$$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$1

For trace-closed finite-volume MPS vectors $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$2, the normalized finite-volume expectations converge to $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$3, with quantitative estimate

$$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$4

For separated local observables $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$5 and $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$6 with $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$7,

$$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$8

Thus thermodynamic limits, boundary stability, and clustering are governed by the same auxiliary product coefficients (Pathirana, 30 Apr 2026).

In random MPS with stationary CPTP transfer cocycles, the right-tail transfer product is exactly a pullback product for the cocycle. Under $$\kappa_{\rm tr}(T)=\frac12\sup\{\|T(E-F)\|_1:\ E,F\text{ orthogonal rank-one projections}\}.$$9 almost surely, there is a unique stationary random boundary state $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$00 with

$$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$01

together with quenched convergence and quenched exponential clustering bounds controlled by the same exponential rate (Pathirana, 30 Apr 2026).

A recurring misconception is that “centered trace-Dobrushin coefficient” is a universally standardized term across the Dobrushin literature. The record is more fragmented. In Gibbs point processes, the standard total-variation Dobrushin matrix is used after discretization, and no separate centered trace-Dobrushin coefficient is introduced (Houdebert et al., 2020). In non-homogeneous Markov-chain CLT papers, the classical Markov–Dobrushin coefficient is retained, while the effective refinement is multi-step or masked—such as the two-step coefficient $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$02 or the restricted quantity $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$03—and the centering lies in centered observables rather than in a separately named coefficient (Veretennikov et al., 2024, Veretennikov et al., 8 Jun 2025). In privacy theory, the standard Dobrushin coefficient remains the total-variation contraction coefficient $$H:=\{X=X^*\in M_d(\mathbb C):\operatorname{Tr}X=0\}$$04, even when sharpened by bounded pointwise maximal leakage constraints (Grosse et al., 14 Jan 2026). In Gibbs-field uniqueness theory, the operative objects are instead the influence matrix, the moment-control matrix, and the resulting contraction matrix (Conache et al., 2015).

The technically precise contemporary usage is therefore narrow: the explicit phrase denotes a trace-norm contraction coefficient on the trace-zero self-adjoint sector of quantum channels, especially for channel products (Pathirana, 30 Apr 2026). At the same time, the broader literature shows that the underlying idea—measuring contraction only after removing the stationary or trivial direction, and doing so in a norm naturally dual to distinguishability—has much older and wider antecedents (Mukhamedov et al., 2020, Gaubert et al., 2013).