Third-Quantized Hartree Approach

• Third-quantized Hartree approach is a method that combines operator-space third quantization with Hartree decoupling to reformulate open interacting bosonic lattice problems.
• The method approximates the quartic Bose–Hubbard interaction by density-dependent onsite shifts, reducing the problem to a solvable quadratic effective Liouvillian.
• It scales polynomially with system size and effectively captures weakly interacting nonequilibrium steady states, including transport and rectification phenomena.

The third-quantized Hartree approach is a self-consistent steady-state method for open interacting bosonic lattices governed by a Lindblad master equation. In its established open-system sense, it combines bosonic third quantization—an operator-space formalism that is exact for quadratic Hamiltonians with linear Lindblad operators—with a Hartree decoupling of the quartic Bose–Hubbard interaction into density-dependent onsite shifts. The resulting effective problem is quadratic, can be solved exactly within third quantization, retains the full infinite bosonic Fock space, and has resource demand that scales polynomially with system size. The method is formulated for weak interactions, nonequilibrium steady states, and transport geometries with boundary driving (Espinoza-Ortiz et al., 2024).

1. Conceptual definition and placement

In the open-quantum-systems literature, third quantization extends the logic of ordinary Fock-space quantization from states in Hilbert space to operators, especially density matrices and observables. Instead of treating only vectors in a many-body Hilbert space, one treats operators on that Hilbert space as elements of an operator space, with superoperators acting as creation- and annihilation-like maps. This is useful because open-system generators act linearly on density operators, and for quadratic or quasi-free models the Liouvillian itself becomes quadratic in these superoperators, allowing exact diagonalization by a non-unitary analogue of the Bogoliubov–de Gennes transformation (Seligman et al., 2010).

Within that framework, the Hartree component enters only when the physical model is interacting. The foundational third-quantization formalism does not derive a Hartree approximation directly; rather, it provides the algebraic machinery needed to replace a non-quadratic Liouvillian by an effective quadratic one and then solve that quadratic problem exactly. The third-quantized Hartree approach for open-system bosonic transport is precisely such a construction: the quartic onsite interaction of the Bose–Hubbard model is replaced by a self-consistent density shift, and the resulting quadratic Lindbladian is handled by third quantization (Seligman et al., 2010).

This places the method between exact quadratic third quantization and generic mean-field treatments. It is more structured than an unspecified density mean field because the effective problem is solved in a fully third-quantized operator-space representation, but it remains an approximation because only Hartree density shifts are retained.

2. Operator-space foundations

The bosonic formalism starts from a many-body Fock space $\mathcal H$ generated by bosonic operators $a_j,a_j^\dagger$ obeying

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$

Third quantization replaces vectors in $\mathcal H$ by operators on $\mathcal H$. For bosons, this requires a dual pair of operator spaces rather than a single globally harmless Hilbert space, because the physical creation and annihilation operators are unbounded. One introduces a space $\mathcal K$ containing trace-class operators such as density matrices and a dual space $\mathcal K'$ containing generally unbounded observables, paired by

$(A|\rho)=\operatorname{tr}(A\rho).$

This dual-space construction is one of the central conceptual features of bosonic third quantization (Seligman et al., 2010).

The basic superoperators are left and right multiplications: $$\hat b^{\mathrm L}|\rho)=|b\rho), \qquad \hat b^{\mathrm R}|\rho)=|\rho b),$$ with adjoint action defined through the trace pairing. From these, the bosonic canonical maps are introduced as

$$\hat a_{0,j}=\hat a_j^{\mathrm L},\qquad \hat a'_{0,j}=\hat a_j^{\dagger\,\mathrm L}-\hat a_j^{\dagger\,\mathrm R},$$

$a_j,a_j^\dagger$0

satisfying the almost-canonical commutation relations

$a_j,a_j^\dagger$1

The qualifier “almost” indicates that $a_j,a_j^\dagger$2 need not be Hilbert adjoints of $a_j,a_j^\dagger$3, although algebraically they behave as canonical conjugates (Seligman et al., 2010).

The construction also defines a left vacuum and a right vacuum,

$a_j,a_j^\dagger$4

and hence a Liouville-Fock basis of operator states. In that basis, quadratic Liouvillians admit normal-mode decomposition, and the non-equilibrium steady state is obtained as the right-vacuum state of the normal master modes. This exact quadratic solvability is the algebraic foundation reused by the Hartree approach (Seligman et al., 2010).

3. Hartree reduction of the open Bose–Hubbard model

The concrete third-quantized Hartree method is formulated for the one-dimensional Bose–Hubbard model on $a_j,a_j^\dagger$5 sites,

$a_j,a_j^\dagger$6

with boundary coupling to Markovian reservoirs through linear jump operators at sites $a_j,a_j^\dagger$7 and $a_j,a_j^\dagger$8. In this setting, bosonic third quantization is exact when the Hamiltonian is quadratic in creation and annihilation operators and the Lindblad operators are linear in those operators. The quartic onsite interaction is therefore the obstruction that must be approximated (Espinoza-Ortiz et al., 2024).

The Hartree step introduces the site density

$a_j,a_j^\dagger$9

and replaces the quartic interaction by

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$0

After dropping the scalar constant proportional to $$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$1, the interaction becomes

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$2

The effective Hamiltonian is therefore

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$3

so the interaction is replaced by a density-dependent onsite energy shift (Espinoza-Ortiz et al., 2024).

Once the Hamiltonian is quadratic, the third-quantized Liouvillian can be written in matrix form as

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$4

with

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$5

and

$$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$6

For the boundary driving used in the transport problem, $$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$7, while $$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$8 and $$[a_j,a_k^\dagger]=\delta_{j,k}, \qquad [a_j,a_k]=[a_j^\dagger,a_k^\dagger]=0.$$9 are diagonal matrices fixed by the reservoir parameters. The nonequilibrium steady state is encoded by the one-body correlation matrix $\mathcal H$0 with entries

$\mathcal H$1

which satisfies the Sylvester equation

$\mathcal H$2

The Hartree closure is therefore a self-consistent prescription for making $\mathcal H$3 depend on $\mathcal H$4 and then solving this quadratic steady-state problem (Espinoza-Ortiz et al., 2024).

4. Self-consistent scheme and observables

The algorithm is an iterative density self-consistency loop. One starts from an initial guess $\mathcal H$5, constructs the effective single-particle Hamiltonian with Hartree diagonal entries,

$\mathcal H$6

forms

$\mathcal H$7

solves

$\mathcal H$8

updates the densities by

$\mathcal H$9

and iterates until

$\mathcal H$0

The converged profile defines the Hartree-renormalized nonequilibrium steady state (Espinoza-Ortiz et al., 2024).

All one-body steady-state observables follow from $\mathcal H$1. The local density is

$\mathcal H$2

and the current operator between sites $\mathcal H$3 is

$\mathcal H$4

so the steady-state current is

$\mathcal H$5

For rectification studies, the ratio between forward and reverse currents is defined as

$\mathcal H$6

with the convention $\mathcal H$7 and hence $\mathcal H$8 (Espinoza-Ortiz et al., 2024).

The computational advantage is structural rather than perturbative. Because the method reduces the interacting problem to matrix equations on $\mathcal H$9 or $\mathcal K$0 objects, it scales polynomially with system size while preserving the full infinite bosonic Fock space of the quadratic Lindblad problem. The paper also notes that, in practice, the steady state can be obtained directly from the reduced Sylvester equation for $\mathcal K$1, so one need not construct the full $\mathcal K$2 matrix $\mathcal K$3 unless desired (Espinoza-Ortiz et al., 2024).

5. Benchmark systems and transport behavior

The method is examined on three one-dimensional Bose–Hubbard transport problems. The first is the uniform chain, with $\mathcal K$4 and $\mathcal K$5. In this case, the steady-state current decreases as $\mathcal K$6 increases, and finite-size scaling converges smoothly. Compared with QuTiP simulations using capped boson numbers for very small systems, the method reproduces the noninteracting limit and captures qualitative trends for $\mathcal K$7, but tends to overestimate the current. The authors repeatedly suggest that the third-quantized Hartree result serves as an upper bound or at least an upper estimate for the steady-state current (Espinoza-Ortiz et al., 2024).

The second benchmark is the interaction-induced diode geometry, in which the hopping is uniform but only one half of the chain is interacting. Forward and reverse bias are realized either by swapping the reservoir occupations or, equivalently, by reversing the interaction profile. The method captures interaction-induced rectification: there is no rectification at $\mathcal K$8, while $\mathcal K$9 decreases with increasing $\mathcal K'$0. Finite-size scaling up to $\mathcal K'$1 converges, and the large-$\mathcal K'$2 behavior displays damped oscillations before saturation. The paper fits the rectification ratio as

$\mathcal K'$3

and interprets the non-monotonic dependence of $\mathcal K'$4 and $\mathcal K'$5 on $\mathcal K'$6 as evidence that rectification is related to mismatch of wave packets across the inhomogeneous interaction interface (Espinoza-Ortiz et al., 2024).

The third benchmark is the SSH–Hubbard model with alternating hoppings and uniform onsite interaction. Forward and reverse configurations are defined by swapping the dimerization pattern rather than the reservoirs. The method finds rectification already at $\mathcal K'$7 and concludes that it is not a bulk band effect, because the SSH bulk bands are invariant under $\mathcal K'$8. Instead, rectification is attributed to different system-reservoir edge couplings after swapping the dimerization pattern. For weak interaction, $\mathcal K'$9, the rectification weakens as $(A|\rho)=\operatorname{tr}(A\rho).$0 increases, and large-$(A|\rho)=\operatorname{tr}(A\rho).$1 results converge rapidly (Espinoza-Ortiz et al., 2024).

Across all three benchmarks, the central empirical pattern is consistent. The third-quantized Hartree approximation captures qualitative transport behavior in the weakly interacting regime, can access system sizes of order $(A|\rho)=\operatorname{tr}(A\rho).$2–$(A|\rho)=\operatorname{tr}(A\rho).$3, and provides finite-size-scaling evidence for convergence toward the thermodynamic limit. Its deviations from exact small-system calculations grow as interactions become stronger, especially once $(A|\rho)=\operatorname{tr}(A\rho).$4 (Espinoza-Ortiz et al., 2024).

6. Limits, misconceptions, and related research programs

The approximation is explicitly a pure density mean-field closure. It keeps Hartree density shifts but does not include exchange or Fock terms, anomalous averages such as $(A|\rho)=\operatorname{tr}(A\rho).$5, pairing channels, or higher connected correlations. It is expected to work best when interactions are weak, steady-state observables are dominated by one-body correlations, and strong quantum correlations beyond density mean field are not essential. The paper associates the weakly interacting regime with

$(A|\rho)=\operatorname{tr}(A\rho).$6

in its units and confines its main claims to steady states under Markovian Lindblad dynamics (Espinoza-Ortiz et al., 2024).

A common misconception is to identify any use of “third quantization” with the open-system operator-space formalism employed here. The term is polysemous. In quantum cosmology and multiverse models, third quantization refers instead to quantization of the wavefunction of the universe, with universes as the quanta; related work includes cubic universe-splitting interactions, time-dependent Wheeler–DeWitt equations, GUP-deformed multiverse quantization, and exact Gaussian doubleverse entanglement models (Faizal, 2014, Balcerzak et al., 2020, Faizal, 2015). In quantum optics, a further distinct usage promotes the quadrature-space wave function of an electromagnetic mode to a field operator, introducing a third-quantized hyperspace of “oscillatons” (Franson, 2021). The third-quantized Hartree approach for bosonic transport belongs specifically to the open-system Liouville-space lineage founded by Prosen and Seligman, not to these cosmological or quadrature-field variants.

Another potential source of confusion is the relation to Hartree theory in equilibrium many-body bosonics. The rigorous derivation of Hartree’s theory for generic mean-field Bose systems proceeds through reduced density matrices, weak quantum de Finetti, and geometric localization in Fock space, with the limiting object described by a measure on one-body states or density matrices (Lewin et al., 2013). That program addresses ground-state energy in the mean-field regime and does not use third quantization in the open-system sense. The overlap is conceptual rather than formal: both frameworks reduce many-body structure to one-body data, but one does so through operator-space Lindbladians and the other through large-$(A|\rho)=\operatorname{tr}(A\rho).$7 marginal limits.

Finally, exact quadratic reformulations continue to sharpen the Gaussian baseline on which a Hartree closure rests. A 2024 reformulation projects quadratic bosonic master equations onto a superoperator coherent-state basis adapted to the $(A|\rho)=\operatorname{tr}(A\rho).$8 representation and shows that the dynamics reduces to a multidimensional complex Ornstein–Uhlenbeck process. This does not develop a Hartree approximation, but it provides an exact covariance and Lyapunov structure for quadratic Lindbladians. A plausible implication is that any third-quantized Hartree or Gaussian closure that succeeds in reducing a nonlinear bosonic Lindbladian to an effective quadratic one should inherit the same OU or Lyapunov organization at the level of one-body correlations (Dupays, 2024).

In this sense, the third-quantized Hartree approach is best understood as a specific weak-interaction, steady-state, one-body-correlator-based approximation scheme inside the broader open-system third-quantization program. Its distinctive feature is not Hartree decoupling alone, but Hartree decoupling followed by exact solution of the resulting quadratic Liouvillian in Liouville-Fock space.