---
title: Two-Component Gross-Neveu Model
url: https://www.emergentmind.com/topics/two-component-gross-neveu-model
type: topic
---

# Two-Component Gross-Neveu Model

Searching arXiv for recent and foundational papers on the two-component Gross–Neveu model.
The two-component Gross–Neveu model denotes a family of interacting fermionic quantum field theories in which the elementary fermion field is represented by a minimal spinor in low spacetime dimension, most commonly a two-component Dirac or Majorana field in \(1+1\) or \(2+1\) dimensions. In its standard form, the model couples \(N\) fermion flavors through a local scalar four-fermion interaction \((\bar\psi\psi)^2\), and is studied both as a paradigmatic asymptotically free theory and as an effective description of symmetry breaking, dynamical mass generation, and bound-state formation. In \(2+1\) dimensions, the two-component representation is especially consequential because the fermion mass term is parity odd, so dynamical mass generation is directly tied to parity breaking [1612.05900]. In \(1+1\) dimensions, the same model and its chiral or flavor-extended variants support exact large-\(N\) analyses, lattice worldline and tensor-network formulations, and nontrivial phase structure at finite temperature and density [1104.0569], [2011.07538], [2606.21246].

## 1. Definition, field content, and representations

The standard \(1+1\)-dimensional Gross–Neveu model with \(N\) Dirac flavors is
\[
\mathcal{L} = \sum_{i=1}^N \bar\psi_i (\gamma_\mu \partial_\mu + m)\psi_i
  - \frac{g^2}{2}\left(\sum_{i=1}^N \bar\psi_i\psi_i\right)^2,
\]
with each Dirac fermion represented by a two-component spinor [1104.0569]. A two-flavor massive version used in real-time simulation is
\[
\mathcal{L}_{GN}
  = \bar{\psi}_{i}(i\gamma^{0}\partial_{t}+i\gamma^{1}\partial_{x}-m)\psi_{i}
    + \frac{1}{2}g^{2}\big(\bar{\psi}_{i}\psi_{i}\big)^{2},
\]
where \(i=1,2\), \(\psi_i\) is a two-component Dirac spinor, and the model has a global \(SU(2)\) flavor symmetry [2011.07538]. In that formulation, the spinor is decomposed as
\[
\psi_i(x)= \begin{pmatrix} \hat a_{x,i}\\ \hat b_{x,i}\end{pmatrix},
\]
so “two-component” refers both to the spinor structure \((\hat a,\hat b)\) and, in the two-flavor case, to the presence of two flavor species [2011.07538].

In \(2+1\) dimensions, the relativistic Gross–Neveu model for \(N\) flavors of two-component fermions is written as
\[
\mathcal{L}_{\text{GN}^{(z=1)}}
 = \sum_{j=1}^N
 \left[ \bar\psi_j\, i\gamma^\mu \partial_\mu \psi_j
  - \frac{g}{2N}(\bar\psi_j\psi_j)^2 \right],
\]
with \(2\times2\) gamma matrices obeying
\[
\{\gamma^\mu,\gamma^\nu\}=2\eta^{\mu\nu},\qquad \eta=(+,-,-)
\]
[1612.05900]. In that setting, a two-component spinor is an irreducible representation of the Lorentz group, and the scalar bilinear \(\bar\psi\psi\) is parity odd, so a dynamically generated mass implies dynamical parity breaking [1612.05900].

A recurrent reformulation replaces Dirac fermions by Majorana fields. In \(1+1\) dimensions, one two-component Dirac fermion equals two two-component Majorana fermions, and the \(N\)-flavor model can therefore be rewritten in terms of \(2N\) Majorana fields with manifest \(O(2N)\) symmetry [1104.0569]. In \(2+1\) dimensions, a \(2N\)-component Dirac field can likewise be decomposed into real Majorana components, exposing an \(\mathrm{SO}(2N)\) structure that organizes mass bilinears into singlet, symmetric-tensor, and adjoint channels [2403.09627], [2406.01681].

## 2. Symmetry structure and mass terms

The defining interaction channel of the canonical model is the scalar four-fermion term \((\bar\psi\psi)^2\), but the symmetry content depends strongly on dimension and representation. In \(2+1\) dimensions with two-component fermions, parity may be defined by
\[
(x_0,x_1,x_2)\to (x_0,-x_1,x_2),\qquad
\psi(x_0,x_1,x_2)\to \gamma^1\psi(x_0,-x_1,x_2),
\]
under which \(\bar\psi\psi\) changes sign [1612.05900]. A nonzero fermion mass therefore breaks parity dynamically. The same parity sensitivity persists in the Hořava–Lifshitz-like generalization, where the spatial kinetic operator is \((i\gamma^i\partial_i)^z\): for odd \(z\) the kinetic term is parity even and the mass term parity odd, whereas for even \(z\) the parity properties differ and no dynamical mass is generated [1612.05900].

In \(1+1\) dimensions, the conventional model has a global \(\mathrm{U}(1)\) charge symmetry and \(\mathrm{SU}(N)\) flavor symmetry, and, in chiral variants, additional axial structure. The generalized Gross–Neveu model with both scalar and pseudoscalar channels,
\[
S = \int \mathrm{d}x\,\mathrm{d}t\left( \sum_{c=1}^N \bar{\psi}_c\,i\slashed{\partial}\,\psi_c
 + \frac{g_x^2}{2N}\left(\sum_c\bar{\psi}_c\psi_c\right)^2
 + \frac{g_y^2}{2N}\left(\sum_c\bar{\psi}_c i\gamma_5\psi_c\right)^2 \right),
\]
has a continuous axial symmetry on the line \(g_x^2=g_y^2\), where the model becomes the chiral Gross–Neveu model [2111.14652]. In that case, the bilinears \(\bar\psi\psi\) and \(\bar\psi i\gamma_5\psi\) form a doublet under axial rotation, and bosonization shows that the infrared theory contains a compact boson whose shift symmetry realizes the continuous chiral symmetry [2111.14652]. The same work emphasizes a mixed ’t Hooft anomaly between the continuous chiral symmetry and charge conservation, implying the existence of a massless mode in the continuum theory [2111.14652].

The O\((2N)\) and \(\mathrm{SO}(2N)\) reformulations refine this symmetry perspective. In the \(1+1\)-dimensional Wilson-fermion loop formulation, the continuum theory
\[
\mathcal{L} = \frac{1}{2}\sum_{i=1}^{2N} \bar\xi_i(\gamma_\mu\partial_\mu + m)\xi_i
  - \frac{g^2}{8}\left(\sum_{i=1}^{2N}\bar\xi_i\xi_i\right)^2
\]
makes the \(O(2N)\) invariance manifest [1104.0569]. In \(2+1\) dimensions, the \(\mathrm{SO}(2N)\) tensor program identifies the canonical Gross–Neveu model as the \(N_f=1\) slice of a broader theory where a real symmetric traceless tensor order parameter unifies all Lorentz-invariant mass-gap orders for \(N\) two-component Dirac fermions except the \(\mathrm{SO}(2N)\)-singlet anomalous quantum Hall state [2406.01681]. This suggests that the “two-component Gross–Neveu model” is not a single universality class once tensor and adjoint channels are admitted, but rather the singlet member of a larger family [2403.09627], [2406.01681].

## 3. Dynamical mass generation and phase structure

The standard nonperturbative phenomenon associated with the model is dynamical mass generation. In the Hořava–Lifshitz-like \(2+1\)-dimensional theory
\[
S = \int dt\,d^2x \; \sum_{j=1}^N \left[ \bar{\psi}_j\big(i\gamma^0\partial_0 + (i\gamma^i\partial_i)^z\big)\psi_j -\frac{g}{2N}(\bar{\psi}_j\psi_j)^2 \right],
\]
the gap equation displays a sharp parity-dependent dichotomy: for even \(z\), dimensional regularization yields only the trivial solution \(m=0\), while for odd \(z\) one obtains
\[
m^{z-2}
  = -\frac{g}{4\pi^{3/2}\,\Gamma\!\left(1+\frac{1}{z}\right)\Gamma\!\left(\frac{1}{2}-\frac{1}{z}\right)},
\]
so a nonzero dynamical mass exists for attractive coupling \(g<0\) and parity is broken dynamically [1612.05900]. At finite temperature, the same theory admits a critical temperature \(T_c\) at which the parity-breaking solution disappears; for \(z=3\) the paper gives an explicit closed form for \(T_c\) in terms of \(m^3\), \(\zeta(1/3)\), and \(\Gamma(1/6)\) [1612.05900].

A distinct nonperturbative treatment based on the Cornwall–Jackiw–Tomboulis effective action finds three different possible mass-generating phases in the massless \(2+1\)-dimensional model, depending on the cutoff dependence chosen for the bare coupling \(G(\Lambda)\) [2111.04539]. At first order in \(G\), the CJT stationarity equations admit: a Dirac mass \(m_D\) phase with \(m_D(g_D+|m_D|)=0\), a Haldane mass \(m_H\) phase with \(m_H(g_H+|m_H|)=0\), and a mixed \((m_3,m_5)\) phase with
\[
m_3\left(g+\sqrt{m_3^2+m_5^2}\right)=0,\qquad
m_5\left(g+\sqrt{m_3^2+m_5^2}\right)=0,
\]
leading to \(\sqrt{m_3^2+m_5^2}=|g|\) for \(g<0\) [2111.04539]. The paper interprets these as three different nontrivial phases distinguished by which discrete symmetries are broken and by which dynamical mass appears [2111.04539].

In \(1+1\) dimensions, finite-density large-\(N\) analysis has recently introduced a different meaning of “two-component”: only a fraction \(\nu=a/N\) of flavors couples to the chemical potential, while the remaining fraction is neutral [2606.21246]. The grand potential in the homogeneous phase becomes
\[
\Psi_{\text{eff}}^{(2)}(\nu;\beta,\mu)
= (1-\nu)\,\Psi_{\text{eff}}^{(1)}(\beta,0) + \nu\,\Psi_{\text{eff}}^{(1)}(\beta,\mu),
\]
and the full phase diagram in \((\mu,T,\nu)\) exhibits a qualitative change at the critical filling fraction
\[
\nu_0 = 0.8271,\qquad \mu_0=0.7202,\qquad T_0=0.2240
\]
[2606.21246]. For \(1\ge \nu\ge \nu_0\), the homogeneous theory has a tricritical point; for \(\nu<\nu_0\), that tricritical point is replaced by a critical end point inside the massive phase [2606.21246]. Once inhomogeneous condensates are included, the phase diagram contains crystal phases bounded by second-order instability lines obtained from stability analyses rather than a full thermal Hartree–Fock solution [2606.21246].

The broadest recent extension of the mass-generation problem is the \(\mathrm{SO}(2N)\) tensor framework in \(2+1\) dimensions. Mean-field theory suggests a first transition at \(g_{c1}\sim 1/N\), where a singlet mass condenses and a discrete symmetry is broken, and a second transition at \(g_{c2}\sim -N g_{c1}\), where \(\mathrm{SO}(2N)\to\mathrm{SO}(N)\times\mathrm{SO}(N)\) [2403.09627]. Renormalization-group analysis of an \(L\)-copy Majorana extension shows that while symmetric-tensor and adjoint-nematic fixed points exist at large \(L\), they lose criticality as \(L\to1\), and at \(L=1\) become equivalent to the Gaussian fixed point, leaving only the standard Gross–Neveu singlet fixed point as genuinely critical [2403.09627]. A complementary Gross–Neveu–Yukawa analysis of the \(\mathrm{SO}(2N)\) symmetric-tensor order parameter finds that for \(N_f=1\), corresponding to the canonical model, the tensor transition is fluctuation-induced first order, and only for \(N_f>N_{f,c2}\approx 2N-2\) does a stable continuous tensor fixed point appear [2406.01681].

## 4. Bound states, solitons, and real-time dynamics

The model’s nonperturbative content is not limited to mass generation; it also includes bound states and solitons. In the \(1+1\)-dimensional O\((2N)\) Gross–Neveu model with Wilson fermions, the fermion-loop formulation rewrites the partition function as a sum over closed, non-backtracking Majorana loops and monomers with positive weights [1104.0569]. In this representation, open worms sample fermion two-point functions, while composite worms sample bound-state correlators such as
\[
\mathcal O(x)=\bar\psi(x)\gamma_5\psi(x)=\bar\xi_1(x)\gamma_5\xi_2(x),
\]
and the corresponding correlators show clean exponential decay, allowing extraction of both fermion and bosonic bound-state masses near the continuum limit [1104.0569]. Because the worm algorithm updates topological sectors and can work directly at the massless point, it provides determinant-free access to the spectrum with essentially no critical slowing down [1104.0569].

A different route to bound-state physics appears in real-time simulations of the two-flavor massive model using matrix product states. On an 11-site lattice with 4 fermionic modes per site, the authors Trotterize the Hamiltonian and simulate up to four particles initially localized on one site [2011.07538]. They observe that when \(m\) is small and \(g\) is large, two- and four-fermion initial states remain localized and their constituent probability densities evolve “synchronistically,” which the paper interprets as dynamical evidence for bound-state behavior [2011.07538]. Flavor symmetry implies that particles of different flavor but the same spin evolve identically, and the Wilson term is required to suppress doubling artifacts [2011.07538]. A plausible implication is that the two-component model furnishes a useful benchmark for nonequilibrium algorithms because the same parameters that favor bound-state formation in continuum studies also leave a visible signature in real-space dynamics.

The richest exact soliton structure in the data arises in the \(1+1\)-dimensional two-flavor O\((2)_L\times O(2)_R\) chiral Gross–Neveu model,
\[
\mathcal L_{O(2)} = \bar\psi\, i\!\not\!\partial\, \psi + \frac{g^2}{2}\big[ (\bar\psi\psi)^2
 + (\bar\psi\,\tau_1\,\psi)^2
 + (\bar\psi\,\tau_3\,\psi)^2
 + (\bar\psi i\gamma_5 \tau_2\,\psi)^2 \big],
\]
whose vacuum manifold consists of two disjoint circles [2302.07660]. In large \(N\), the model supports a massless Goldstone mode and three massive mesons of mass \(2m\), as well as explicit kink, baryon, and breather solutions [2302.07660]. Via a Majorana decomposition, this model is dual to the “perfect Gross–Neveu model,” a one-flavor theory with both chiral and Cooper-pairing channels and full Pauli–Gürsey symmetry, itself dual to the Zakharov–Mikhailov model [2302.07660]. This demonstrates that two-component Gross–Neveu variants can interpolate between conventional particle–hole condensation and particle–particle pairing while remaining integrable in large \(N\) [2302.07660].

## 5. Lattice formulations and numerical methods

Lattice formulations of the two-component Gross–Neveu model differ substantially in their fermion discretization and in the observables they make most accessible. Wilson fermions provide a direct path to loop/worldline algorithms. In the O\((2N)\) Majorana formulation, the lattice action
\[
S = \frac{1}{2}\sum_x \sum_{i=1}^{2N} \varphi \,\bar\xi_i(x)\xi_i(x)
  - \frac{g^2}{8}\sum_x\left(\sum_{i=1}^{2N}\bar\xi_i(x)\xi_i(x)\right)^2
  - \sum_{i=1}^{2N}\sum_{x,\mu} \bar\xi_i(x) P(\mu)\,\xi_i(x+\hat\mu)
\]
with \(\varphi=2+m\) and \(P(\pm\mu)=\frac12(1\mp\gamma_\mu)\) yields positive loop weights and a worm algorithm that samples both fluctuating topological sectors and bound-state correlators [1104.0569]. Partition-function combinations such as \(Z^{0100\ldots}\) vanish at the critical point, allowing the critical Wilson mass to be located numerically by a zero crossing [1104.0569].

Borici–Creutz fermions provide a minimally doubled, chirally invariant alternative. In \(1+1\) dimensions, the Gross–Neveu action with Borici–Creutz fermions is
\[
\begin{aligned}
S_{BC}&=\sum_{n}\Big[\frac{1}{2}\sum_{\mu}\bar{\psi}_n\gamma_{\mu}(\psi_{n+\mu}-\psi_{n-\mu})
-\frac{i}{2}\sum_{\mu}\bar{\psi}_n(\Gamma-\gamma_{\mu})(2\psi_{n}-\psi_{n+\mu}-\psi_{n-\mu}) \\
&\hspace{2em} +i(c_{3}-2)\bar{\psi}_{n}\Gamma\psi_{n}
+m\bar{\psi}_{n}\psi_{n}
-\frac{g^{2}}{2N}\Big[(\bar{\psi}_{n}\psi_{n})^{2}+(\bar\psi_{n}i\Gamma\psi_{n})^{2}\Big]\Big],
\end{aligned}
\]
with \(\Gamma=\frac12(\gamma_1+\gamma_2)\) in 2D [1409.7999]. Strong-coupling analysis and hybrid Monte Carlo both find a second-order chiral phase transition, with a critical value \(\beta_c\approx0.4\) in the massless limit [1409.7999]. This formulation reproduces the qualitative chiral phase structure of the continuum model while introducing an additional \((\bar\psi i\Gamma\psi)^2\) channel tied to the special Borici–Creutz direction [1409.7999].

Staggered-fermion formulations make it possible to study competition between ordinary bilinear mass generation and more exotic symmetric mass generation. A \(2+1\)-dimensional lattice model with two flavors of massless staggered fermions, a nearest-neighbor current–current interaction \(U\), and an on-site four-fermion interaction \(U'\),
\[
S = S_0
-U \sum_{\langle x,y \rangle} \left( \bar u_x u_x \,\bar u_y u_y + \bar d_x d_x \,\bar d_y d_y \right)
-U' \sum_{x} \left( \bar u_x u_x \,\bar d_x d_x \right),
\]
is free of sign problems in the fermion bag representation [2111.15134]. Based on earlier limits, the authors expect three phases: a PMW phase with massless fermions, an FM phase where fermions become massive through spontaneous symmetry breaking and a bilinear condensate, and a PMS phase where fermions are massive without any bilinear condensate [2111.15134]. This suggests that even in a two-component Gross–Neveu–Thirring setting, the bilinear condensate is not the only possible mass-generating mechanism.

Tensor-network methods furnish a complementary Hamiltonian approach. The two-flavor lattice Hamiltonian used in real-time MPS simulation contains kinetic, Wilson, mass, and on-site interaction terms and acts on \(4n_x\) fermionic modes, encoded as qubits via Jordan–Wigner [2011.07538]. For \(n_x=11\), \(\Delta t=0.1\), and \(n_t=150\), the full state would naively live in a \(2^{45}\)-dimensional Hilbert space, but MPS compression keeps the simulation tractable, with reported MPS errors below \(10^{-9}\) for the few-particle states considered [2011.07538]. The same work quotes a Trotter error of order \((\Delta t)^2\sim 0.01\), highlighting the balance between circuit fidelity and tensor-network compression in explicit time evolution [2011.07538].

## 6. Modern extensions and related universality classes

Several recent developments situate the two-component Gross–Neveu model within a broader web of fixed points and universality classes. In \(2+1\) dimensions, the bosonized Gross–Neveu–Yukawa theory has been pushed to three loops in \(4-\epsilon\) dimensions for general flavor number \(N\), yielding explicit \(\epsilon\)-expansions for \(\nu^{-1}\), \(\eta_\psi\), and \(\eta_\phi\) at the chiral Ising fixed point [1703.08801]. For \(N=1\), the paper quotes
\[
\nu^{-1}\approx 2 - 0.835\,\epsilon - 0.00571\,\epsilon^2 - 0.0603\,\epsilon^3,\quad
\eta_\psi\approx 0.1\,\epsilon + 0.0102\,\epsilon^2 - 0.033\,\epsilon^3,\quad
\eta_\phi\approx 0.4\,\epsilon + 0.102\,\epsilon^2 - 0.0632\,\epsilon^3,
\]
and for \(N=2\),
\[
\nu^{-1}\approx 2 - 0.952\,\epsilon + 0.00723\,\epsilon^2 - 0.0949\,\epsilon^3,\quad
\eta_\psi\approx 0.0714\,\epsilon - 0.00671\,\epsilon^2 - 0.0243\,\epsilon^3,\quad
\eta_\phi\approx 0.571\,\epsilon + 0.124\,\epsilon^2 - 0.0278\,\epsilon^3
\]
[1703.08801]. In the special \(N=1/4\) case, corresponding to a single-component fermion in the paper’s conventions, the exponents satisfy emergent super-scaling relations order by order up to three loops, consistent with \(\mathcal N=1\) supersymmetry [1703.08801].

At the lower-critical end, the fate of the non-supersymmetric Gross–Neveu–Yukawa fixed point with a two-component Majorana fermion continued to two dimensions has been analyzed through fermionic minimal models [2212.06342]. Under the assumptions that the fixed point is a fermionic minimal model with a chiral \(\mathbb Z_2\) symmetry and just two relevant singlet operators, only four candidates survive; matching topological defect line spin content under the assumed flow to the supersymmetric fermionic tricritical Ising model rules out two of them, leaving the fermionic \((11,4)\) and fermionic \((E_6,A_{10})\) models [2212.06342]. An additional double-braiding constraint favors the non-unitary fermionic \((11,4)\) model [2212.06342]. This suggests that the two-dimensional continuation of the non-supersymmetric two-component Majorana GNY fixed point may be non-unitary, even though the higher-dimensional parent theory is treated perturbatively.

The \(\mathrm{SO}(2N)\) generalization provides a final unifying viewpoint. Rewriting the canonical \(2+1\)-dimensional Gross–Neveu interaction in terms of a single \(4N\)-component Majorana field exposes three quartic channels: a singlet scalar, a symmetric tensor, and an adjoint nematic [2403.09627]. Extending the theory to \(L>1\) copies reveals three corresponding fixed points at large \(L\), but only the singlet Gross–Neveu fixed point remains physically critical at \(L=1\) [2403.09627]. A distinct \(2+\epsilon\) analysis of an \(\mathrm{SO}(2N)\)-symmetric, Fierz-complete Majorana theory reaches a closely related conclusion: the Gross–Neveu–Ising fixed point remains critical for all \(N_f\ge1\), whereas the symmetric-tensor fixed point loses criticality below
\[
N_{f,c}^{\mathrm{ST}(N)} \approx 0.56 + 1.48\,N + \mathcal O(\epsilon),
\]
so for the physical \(N_f=1\) case only the singlet Gross–Neveu–Ising transition survives as a genuine continuous transition [2510.23725]. This reinforces a common modern reading: the canonical two-component Gross–Neveu model is robustly associated with singlet mass generation, while more elaborate tensor or nematic channels are generically fluctuation-destabilized unless the fermion flavor content is enlarged [2403.09627], [2406.01681], [2510.23725].

Source: https://www.emergentmind.com/topics/two-component-gross-neveu-model