--- title: Spin-SYK Variants & Infrared Dynamics url: https://www.emergentmind.com/topics/spin-syk-model type: topic --- # Spin-SYK Variants & Infrared Dynamics Searching arXiv for recent and foundational papers on Spin-SYK and related variants. The term **Spin-SYK model** is used in the literature for several closely related extensions of the Sachdev–Ye–Kitaev framework in which spin degrees of freedom, spin symmetry, or spin-sector observables are central. In one usage, Majorana fermions are replaced by Pauli spins or hard-core bosons, yielding all-to-all random spin Hamiltonians such as the SpinXY4 model [2309.15349]. In another, the local variables are SU($d$) generators rather than fermions, leading to “qudit SYK” and other two-local generalizations [2505.09900]. A different line of work studies spinful, time-reversal-invariant SYK models of complex fermions with SU(2) symmetry and pairing channels [2002.11757], while the low-energy theory of SYK models with continuous global symmetry identifies the spin case with $G=\mathrm{SU}(2)$ and an infrared factorization into a Schwarzian mode and a rigid rotor on the group manifold [1901.05666]. Yet another usage concerns “higher-spin SYK,” where the singlet bilinear tower realizes $hs[\lambda]$ or $ehs[\lambda]$ and admits higher-spin BF or Poisson-sigma-model bulk duals [2509.19964]. The label therefore does not denote a single microscopic Hamiltonian, but a family of SYK-type constructions sharing random, strongly interacting, and often maximally chaotic low-energy dynamics. ## 1. Terminological scope and principal variants The literature does not use **Spin-SYK** in a single uniform way. The most direct meaning is a literal spin replacement of the fermionic SYK variables: the SpinXY4 model replaces Majoranas by local Pauli operators, while preserving all-to-all random 4-local interactions [2309.15349]. A second meaning is a **spinful SYK** model, where the microscopic fermions carry SU(2) spin and the low-energy theory includes both Schwarzian and spin-rotor sectors [1901.05666]. A third meaning is **higher-spin SYK**, in which the relevant “spin” is not microscopic SU(2) but an infinite tower of singlet bilinears organized by a higher-spin algebra and described holographically by higher-spin extensions of Jackiw–Teitelboim gravity [2509.19964]. A fourth meaning is phenomenological: SYK-like scaling and operator growth in the spin sector of correlated electron models, especially the spin-freezing crossover of multi-orbital Hubbard systems [1812.04217]. | Usage of “Spin-SYK” | Degrees of freedom | Defining structure | |---|---|---| | Pauli-spin / hard-core-boson model | Local $\sigma^{x,y}$ operators | All-to-all random 4-local spin interactions | | SU($d$) or qudit SYK | SU($d$) generators $T_{i,\alpha}$ | Random $q$-local couplings, including two-local cases | | Spinful fermionic SYK | Complex spin-$1/2$ fermions $c_{i\sigma}$ | SU(2)-invariant random interactions, pairing channels | | Higher-spin SYK | Bilinear singlet tower | $hs[\lambda]/ehs[\lambda]$ symmetry, BF/PSM bulk dual | | Spin-sector SYK phenomenology | Hubbard-Hund local spin observables | $\Sigma(\omega)\sim \sqrt{\omega}$ and SYK-like OTOCs | This plurality is not merely terminological. It reflects distinct research programs: quantum simulation with spins or qudits, infrared effective theory with global symmetry, holographic higher-spin generalizations, and non-random condensed-matter realizations of SYK-like spin dynamics. ## 2. Microscopic constructions In the Pauli-spin realization SpinXY4, the operators are defined by $$ \hat O_{2j-1}=\hat \sigma_{j,x}, \qquad \hat O_{2j}=\hat \sigma_{j,y}, $$ and the Hamiltonian is an all-to-all random 4-local sum of terms $J_{abcd}\, i^{\eta_{abcd}} \hat O_a \hat O_b \hat O_c \hat O_d$, where $\eta_{abcd}$ counts how many sites contribute both $x$ and $y$ components and supplies the Hermiticity factor. The random couplings are i.i.d. standard Gaussian, and the normalization is chosen so that energy, entropy, and characteristic time scales obey the standard SYK large-$N$ scaling. The model admits a hard-core-boson interpretation because $\hat O_{2a-1}\pm i\hat O_{2a}$ act as creation and annihilation operators with on-site exclusion [2309.15349]. The SU($d$) generalization, or **qudit SYK**, replaces Pauli matrices by generators $\tau_\alpha$ of SU($d$), normalized by $$ \mathrm{Tr}(\tau_\alpha \tau_\beta)=2\delta_{\alpha\beta}, $$ with site operators $T_{i,\alpha}$. Its $q$-local Hamiltonian is $$ H_{\mathrm{qudit}}=\sum_{1\le i_1<\cdotsp+1}J_{pq}(\chi_p\chi_{p+1})(\chi_q\chi_{q+1}), $$ with variance $E[J_{pq}^2]=(N-1)/(2N^2)$ chosen to match the energy variance of the original $q=4$ SYK model [2505.09900]. A distinct microscopic construction is the **spin-1/2, time-reversal-invariant SYK model** built from complex fermions $c_{i\sigma}$. Its random interaction is $$ H_{\mathrm{SYK}}=\frac{1}{2}\sum_{ijkl;\sigma\sigma'}J_{ij;kl} \big[ c_{i\sigma}^\dagger c_{j\sigma'}^\dagger c_{k\sigma'} c_{l\sigma} + c_{l\sigma}^\dagger c_{k\sigma'}^\dagger c_{j\sigma'} c_{i\sigma} \big], $$ with real couplings obeying $$ J_{ij;kl}=-J_{ji;kl}=-J_{ij;lk}=J_{lk;ji}, \qquad \langle J_{ij;kl}^2\rangle=\frac{J^2}{(4N)^3}. $$ This model preserves SU(2) and time reversal and becomes the basis for pairing, pseudogap, and superconducting extensions when supplemented by attractive interactions such as a negative-$U$ Hubbard term [2002.11757]. ## 3. Infrared structure, collective fields, and holographic descriptions Across these variants, the low-energy description repeatedly reduces to Schwarzian dynamics plus additional soft sectors. In standard large-$N$ SYK-like treatments one introduces bilocal fields $G(\tau,\tau')$ and $\Sigma(\tau,\tau')$ obeying Schwinger–Dyson equations $$ G=(\partial_\tau-\Sigma)^{-1}, \qquad \Sigma(\tau,\tau')=J^2[G(\tau,\tau')]^{q-1}, $$ with an emergent reparametrization mode governed by $$ S_{\mathrm{Sch}}=-C\int du\,\{f(u),u\}, $$ where $C\sim N\alpha/J$. For a bosonic operator $O(t)$ of scaling dimension $\Delta$, the conformal finite-temperature correlator takes the form $$ G_0(t_1,t_2)=\Big[\frac{\pi}{\beta \sin(\pi t_{12}/\beta)}\Big]^{2\Delta}. $$ In the “spin-SYK” setting of spin-like composite operators, Schwarzian loop corrections reshape the retarded correlator and density of states and generate explicit formulas for local susceptibilities and higher-point OTOCs [1804.06164]. For SYK models with continuous global symmetry $G$, the infrared effective action factorizes as $$ S_{\mathrm{eff}}=S_{\mathrm{Schwarzian}}+S_{\mathrm{group}}, $$ where the group sector is a free particle on the group manifold, $$ S_{\mathrm{group}}[g]=-\frac{K\pi}{2\beta}\int_0^{2\pi} d\tau\, \mathrm{Tr}\big[(g^{-1}\partial_\tau g)^2\big]. $$ For the spin case $G=\mathrm{SU}(2)$, the group partition function is $$ Z_G^{SU(2)}(\beta)=\sum_{J=0,1/2,1,\ldots}(2J+1)^2 e^{-\beta J(J+1)/K}. $$ This establishes the infrared spin sector as a rigid rotor on SU(2), with energies $E_J=J(J+1)/K$ and a modified low-temperature thermodynamics and spectral form factor. The reparametrization mode still controls maximal chaos, and the global-symmetry modes do not increase the Lyapunov exponent beyond $\lambda_L=2\pi/\beta$ [1901.05666]. The Pauli-spin SpinXY4 model has a different but related collective-field structure. After disorder averaging, the real-time path integral retains only $S_N$ permutation symmetry rather than finite-coupling $O(N)$. The large-$N$ description closes on an infinite tower of multi-local invariants $$ \Phi_L(t_1,\ldots,t_L;t_1',\ldots,t_L') = \frac{1}{N}\sum_{i=1}^N S_i^+(t_1)\cdots S_i^+(t_L) S_i^-(t_1')\cdots S_i^-(t_L'). $$ In the strong-coupling limit an emergent $O(N)$ symmetry appears, the theory reduces to a symmetric bilocal field $\phi(X,Y)$, and the collective action becomes $$ A_{\mathrm{col}}[\phi]=-N\,\mathrm{Tr}\ln\phi-\frac{N}{4}\sum_{X,Y}[\phi(X,Y)]^4. $$ The saddle equation is $$ \sum_Z [\phi_0(X,Z)]^3\phi_0(Z,Y)=-\delta_{X,Y}, $$ which is the bosonic-spin analogue of the SYK bilocal formalism [2309.15349]. In higher-spin SYK, the bilinear singlet tower $$ O_n(\tau)=\chi^\dagger(\tau)\cdot(\overleftrightarrow{\partial_\tau})^n \chi(\tau), \qquad n=0,1,2,\ldots $$ has scaling dimensions $\Delta_n=2\Delta+n$, and on the generalized free line with $\Delta=(1-\lambda)/2$ one obtains $\Delta_n=n+1-\lambda$. The bulk dual replaces ordinary JT gravity by a higher-spin BF or Poisson sigma model based on $hs[\lambda]$ or $ehs[\lambda]$. Two realizations are emphasized: **Model A**, built from a deformation of $U(\mathfrak{sl}(2,\mathbb{R}))\rtimes \mathbb{Z}_2$ and formulated as a BF theory with a Lax pair, and **Model B**, a perturbatively local Poisson sigma model derived from a cyclic $A_\infty$ structure. In both cases the scalar tower satisfies $$ m_n^2=(n-\lambda+1)(n-\lambda), $$ and the boundary dynamics is a higher-spin generalization of the Schwarzian action [2509.19964]. Closely related bulk-side analyses derive generalized Schwarzian actions for Yang–Mills and $SL(N)$ higher-spin extensions directly from BF-formulated dilaton gravity and the Lee–Wald–Zoupas symplectic structure [1802.01562]. ## 4. Chaos, level statistics, and the spin-glass question The Pauli-spin SpinXY4 model is strongly chaotic in the random-matrix sense. The operator $$ \hat \Gamma=\hat \sigma_{1,z}\otimes\cdots\otimes \hat \sigma_{N_{\mathrm{spin}},z} $$ commutes with the Hamiltonian, so parity is conserved and the Hilbert space splits into two blocks of dimension $2^{N_{\mathrm{spin}}-1}$. When same-site $x/y$ pairs are allowed, the finite-$N$ random-matrix universality class is GUE for any $N_{\mathrm{spin}}$. Numerically, the density of states over the bulk is nearly indistinguishable from SYK4 up to $N_{\mathrm{Maj}}=34$, the nearest-neighbor spacing ratios agree with GUE except for a few of the lowest levels, and the spectral form factor exhibits long ramps very close to those of SYK4. Two-point functions distinguish parity-flipping and parity-preserving operators: $G_{x,y}(t)$ behaves like the SYK Majorana correlator and lacks a ramp/plateau, whereas $G_z(t)$ shows a ramp/plateau because it probes a single parity sector. Edwards–Anderson diagnostics suggest that a small number of low-energy states may show spin-glass-like tendencies, but over most of the spectrum the model is ergodic and RMT-like [2309.15349]. The SU($d$) and two-local modifications reach a similar conclusion. For qudit SYK with $d=3,4,5,6$ at $q=2$ and $d=3$ at $q=3$, the spacing distribution $P(s)$ converges to the GUE Wigner surmise as $L$ increases, the adjacent-gap ratio concentrates near the GUE value $0.59975$, and the spectral form factor shows a clear dip–ramp–plateau. For the overlapping-clusters model $$ H=\sum_p\sum_{q>p+1}J_{pq}(\chi_p\chi_{p+1})(\chi_q\chi_{q+1}), $$ the expected random-matrix ensemble depends on $N \bmod 8$: GOE for $N \bmod 8=0,2,6$ and GUE for $N \bmod 8=4$ after sector decomposition. Its density of states has soft edges, but the bulk still displays spectral rigidity and Wigner–Dyson statistics over a wide energy window [2505.09900]. A recurring misconception is that SYK’s strong disorder should generically induce a spin-glass phase. For the $q=4$ Majorana SYK model, this was explicitly tested and rejected. Replica off-diagonal condensation is controlled by a replicon mass $m_{ab}^2$; a naive nearly-conformal estimate suggests an exponentially small critical temperature, but for $q=4$ this lies outside the conformal regime, and the correct Schwarzian-regime computation yields $$ m_{ab}^2=\frac{4N-c_2\log^2 N}{(\beta J)^2}, $$ with $c_2>0$, so the replica-symmetric saddle remains stable at large $N$. Numerically, for $N$ up to $46$ Majoranas the edge statistics agree with RMT rather than Poisson, the spectral form factor shows a ramp and plateau, and the ground-state energy distribution is Gaussian rather than Tracy–Widom. The conclusion is that the $q=4$ Majorana SYK model remains paramagnetic and ergodic, with no spin-glass phase [1806.10145]. ## 5. Spin observables, susceptibilities, and pairing instabilities Spin-sensitive observables in SYK-like models can be formulated through bosonic composite operators $O(t)$ of infrared scaling dimension $\Delta$. In that framework, the retarded two-point function receives explicit Schwarzian loop corrections, and the density of states $$ \rho(\omega)=-\frac{1}{\pi}\,\mathrm{Im}\,G^R(\omega) $$ undergoes a transfer of spectral weight from the low-frequency quasiparticle peak to finite-frequency sidebands identified as Hubbard bands. The same Schwarzian theory gives an exponentially growing crossed OTOC, $$ G^{(4)}_{VWVW}(t)\approx -4^{-2\Delta}\frac{\Delta^2}{C}\frac{1}{\beta}e^{\lambda_L t}, \qquad \lambda_L=\frac{2\pi}{\beta}, $$ and explicit local spin susceptibilities. At finite temperature, $$ \chi_{\mathrm{loc}}^{(2)}(\omega) = \Big(\frac{\beta}{2\pi}\Big)^{1-4\Delta} \Gamma(1-4\Delta) \frac{\Gamma\!\left(2\Delta-\frac{i\omega\beta}{2\pi}\right)} {\Gamma\!\left(1-2\Delta-\frac{i\omega\beta}{2\pi}\right)}. $$ For the marginal non-Fermi-liquid case $\Delta=1/4$, $$ \mathrm{Im}\,\chi_{\mathrm{loc}}^{(2)T}(\omega)=\frac{\pi}{2}\tanh\!\Big(\frac{\omega}{2T}\Big), $$ while higher derivatives give compact $\mathrm{sech}^2$ and $\tanh\,\mathrm{sech}^2$ expressions relevant to dynamical susceptibilities [1804.06164]. In the spin-1/2, time-reversal-invariant fermionic SYK model with onsite attractive interaction, the basic pairing operators are $$ b_i^\dagger=c_{i\uparrow}^\dagger c_{i\downarrow}^\dagger, \qquad \rho_{ij}=\langle b_i^\dagger b_j\rangle. $$ Off-diagonal long-range order is diagnosed by an extensive largest eigenvalue $\lambda_0$ of $\rho_{ij}$. Exact diagonalization at half-filling for $N=12$ finds a finite critical attraction $U_c\approx 0.24J$ below which there is neither ODLRO nor a many-body gap beyond the small finite-size SYK gap; instead the system enters a pseudogap regime dominated by phase fluctuations. Those fluctuations are described by a quantum Kuramoto action, $$ S[\{\phi_i(\tau)\}] = \int d\tau \Big[ \frac{m}{2}\sum_i (\partial_\tau \phi_i)^2 - \frac{g}{N}\sum_{i0$, the pure Kitaev limit has a ground-state degeneracy scaling as $2^{N/2}$, first-order matrix elements of the $\Gamma$ perturbation vanish inside the zero-mode manifold, and in the anisotropic regime $K_x\gg K_y>0$ one can project to an effective quartic Hamiltonian $$ H_{\mathrm{eff}} \approx -\frac{1}{\Delta E}P_0 H_1^2 P_0 = \sum_{i