WorksPhysics and Astronomy
Study of the frustrated SU(3) Heisenberg chain with SU(3)-symmetric matrix product states
Abstract
We study the SU(3) Heisenberg chain with nearest- and next-nearest-neighbor antiferromagnetic exchange and . Field theory predicts that the trimerized valence-bond solid of this chain has a finite extent in , followed by a gapless SU(3) phase, but earlier numerical estimates of the upper boundary range from (exact diagonalization) to (iDMRG). We use the variational uniform matrix product state (VUMPS) algorithm with the full non-Abelian SU(3) symmetry imposed on the virtual bonds, reaching bond dimensions up to , and locate the phase boundaries by finite-entanglement scaling. Resolving the transfer-matrix spectrum and the entanglement spectrum by SU(3) irreducible representation, we find that the trimerized phase extends to and end below , excluding a boundary near . At larger our numerical analysis of the effective central charge is consistent with the predicted massless flow from SU(3) to SU(3).
Introduction
Lattice models with SU() symmetry describe Mott insulators of alkaline-earth-like atoms with populated nuclear-spin states [1, 2], which have been realized with and in optical lattices [3, 4, 5, 6]. For two spins in the fundamental representation cannot form a singlet; the smallest singlet involves sites. Frustration therefore favors -site valence-bond solids rather than dimers.
The simplest frustrated model of this kind is the – SU(3) Heisenberg chain [Eq. (1)]. At zero next-nearest-neighbor coupling (), it is the integrable Sutherland chain [7], a gapless SU(3) Wess–Zumino–Witten (WZW) critical point with central charge . Increasing drives a marginal current–current coupling through zero and opens a gap into a threefold-degenerate trimerized phase at [8, 9, 10]. The fate of this phase at large was clarified by Herviou, Capponi and Lecheminant (HCL) [10]: viewing the chain as a zigzag ladder of two SU(3) chains coupled by , they showed that for odd , the inter-chain coupling gaps a parafermion sector and leaves an SU() theory perturbed by its adjoint field, which flows without a gap to SU() [11]. The trimerized phase thus has a finite extent, and a second gapless SU(3) phase appears at large .
The location of the upper boundary for the trimerized phase and the nature of the large- phase are not settled numerically. Exact diagonalization (ED) on rings of up to 27 sites gives with strong finite-size effects, while iDMRG with U(1) × U(1) color conservation and bond dimension shows trimerization up to , beyond which the data jump at a -dependent coupling [10]; an earlier study placed the boundary near 3.5 [8]. At large , HCL obtained an effective central charge , close to the SU(3) value , and argued that full SU(3) symmetry would be needed to resolve the flow to , in analogy with the SU(3) chain in the symmetric representation [12, 13].
Here we address both questions with variational uniform matrix product states (VUMPS) [14, 15] with exact SU(3) symmetry on the virtual bonds, at bond dimensions up to [about 540 SU(3) multiplets]. Our main results are:
An injective SU(3)-symmetric MPS with the fundamental representation on every site cannot be invariant under one-site translation; finite- states therefore break translation and generically show some trimerization, so that the phase boundaries have to be located by finite-entanglement scaling (§ 2).
At the lower transition, the extrapolated trimerization becomes positive at , once states up to are employed, consistent with the earlier ED estimate. By contrast, a fit of the essential-singularity form expected for this SU(3)-symmetric marginal transition cannot locate at reachable bond dimensions (§ 2.4).
From the bond-dimension scaling of the octet correlation length, the entanglement entropy and the trimerization, the trimerized phase extends to ; where it ends is not resolved at the bond dimensions reached (§ 2.5).
Above , for , the effective central charge lies below , and where the crossover length is shortest (moderate ), it keeps decreasing with (Sec. VI).
The jumps of the iDMRG data of Ref. [10] are crossings, at fixed bond dimension, to states outside the SU(3)-symmetric ansatz, which we reproduce with an Abelian iDMRG approach; no SU(3)-symmetric state of either entanglement structure reproduces them (Sec. VII).
After the model and the method (§ 2) we give an overview of the phase diagram (Sec. III), determine the lower and the upper boundary of the trimerized phase (§ 2.4 and § 2.5), discuss the large- regime (Sec. VI), and compare with previous work (Sec. VII).
Model and method
Model
We consider
where each site carries the fundamental representation of SU(3) and exchanges the states of sites and . With SU(3) generators normalized as one has , so Eq. (1) differs from the spin form used in Ref. [10] only by a factor of two, and coincides with . We set . Drawing the chain as a zigzag ladder (Figure 1), the bonds form two legs and the bonds alternate between “rung” and “diagonal” bonds. Two sites cannot form an SU(3) singlet, since ; the smallest singlet involves three sites, which is why frustration favors period-3 (trimer) rather than bipartite order.

Figure 1. Zigzag chain geometry and the SU(3) permutation operator. Sites (circles) carry one of three colors (red = 1, green = 2, blue = 3), illustrating the period-3 pattern of the trimerized phase. Solid lines are nearest-neighbor () bonds; within the zigzag, two geometrically distinct subtypes arise – “rung” bonds (bottom→top, carrying energy ) and “diagonal” bonds (top→next-bottom, ) – which can have different expectation values whenever translational symmetry is broken. Dashed lines are next-nearest-neighbor () bonds running along each row (), which introduce geometric frustration. The bracketed region is one unit cell ( sites). Inset: the permutation operator exchanges the SU(3) colors of sites and . It plays the same role as in SU(2) but has no two-site singlet eigenstate, so color-neutral clusters require three sites.
At , Eq. (1) is the Sutherland chain, with exact ground-state energy per site [7] and an SU(3) critical point with . At the two legs decouple into two such chains (). For any finite , the inter-leg coupling is relevant: it gaps a parafermion sector () and leaves SU(3) () perturbed by its adjoint primary, which flows without a gap to SU(3) () [10, 11]. Here for SU(), i.e. for and for ; the parafermion sector accounts for .
We use a unit cell of six sites (three rungs), commensurate with both the period-3 order and the two-leg structure, as in Ref. [10]. All states we obtain have a period of three sites (the bond energies repeat after three sites), thus a three-site cell would describe the same states; the six-site cell, in addition, admits period-2 and period-6 patterns, which do not occur.
SU(3)-symmetric uniform MPS
Ground states are obtained with the variational uniform matrix-product-state algorithm (VUMPS) [14, 15], with the SU(3) symmetry imposed exactly on the virtual spaces [16, 17, 18, 19], as in non-Abelian symmetric DMRG and tensor-network methods [20, 21]: every tensor is stored as reduced blocks between multiplets, with the Clebsch–Gordan structure fixed by the symmetry. Each virtual space is a direct sum of SU(3) irreducible representations (irreps) with multiplicities; the bond dimension counts states, so an irrep with multiplicity contributes . At the states carry about 420 multiplets ( states per multiplet), and the cost is set by the multiplet structure rather than by . Because each site adds one box to the Young diagram, the triality (number of boxes mod 3) of the virtual irreps advances by one from bond to bond: every bond carries irreps of a single triality class, (), () or (their conjugates), cycling through the unit cell. The irreps and their multiplicities are not fixed in advance: they are selected by the singular values, and irreps up to dimension 64 occur in our states. The truncation keeps or discards whole multiplets, ranked by their Schmidt values, so the state is exactly SU(3)-symmetric at every step; the triality of the irreps on each bond is fixed by the fusion rules and is not chosen by hand.
The bond dimension is increased in steps of a factor 1.2. Each step starts with a two-site infinite DMRG sweep that truncates to the new ; this is the step in which the bond spaces are chosen, i.e., which irreps occur on each bond and with which multiplicities, by keeping the multiplets with the largest Schmidt values. VUMPS then optimizes the tensors at fixed bond spaces until the gradient norm
falls below (at most 1000 iterations). Here and are the left-orthonormal and center-site tensors of site of the cell, and is the effective Hamiltonian acting on ; is the norm of the energy gradient projected onto the tangent space of the MPS manifold [14, 15] and vanishes at a variational optimum.
No single-site translation invariance at finite
The fundamental representation is a projective representation of PSU(3)SU(3)/: the center element , , acts as the phase on each site. Consider a single-site uniform MPS with this symmetry. Each site adds one box to the Young diagram, so maps virtual irreps of triality to triality ; with on the triality- blocks of the virtual space, . If is the fixed point of the transfer map , then ; since is invertible, , and is a second eigenvalue of modulus one. An injective MPS has a unique eigenvalue of modulus one, so an injective SU(3)-symmetric MPS cannot be invariant under one-site translation. This is the matrix-product-state form of the Lieb–Schultz–Mattis theorem [22, 23], shown for SU(2) with half-integer spin in Ref. [24]; a complete derivation is given in Appendix C.
A translation-invariant SU(3)-symmetric MPS exists only in non-injective form, as the equal-weight superposition of the three translates of a period-3 injective state; its correlations do not cluster and it carries no information beyond its components. In the SU(3)-symmetric representation, the translation breaking of an injective state is visible directly in the entanglement spectrum. On cuts with every irrep has dimension divisible by three, so every Schmidt value is at least threefold degenerate, while the cuts contain the singlet sector with non-degenerate values. The spectra of neighboring cuts therefore differ, and every finite- state breaks single-site translation: the bond energies are modulated with a period of three sites, with in all our states, also in a gapless phase (at , at ). The distinction between the gapless and the trimerized phases is then a matter of -scaling: , and saturate in the trimerized phase, while in a gapless phase and , grow without bound.
Observables
Correlation length. The transfer matrix of the six-site cell, , is block diagonal in the SU(3) irrep of its channel (only triality-zero irreps occur). Let be its eigenvalues in sector , ordered by modulus. For a normalized injective MPS the largest eigenvalue is , in the singlet sector; it corresponds to the norm of the state and carries no correlations. We therefore define, in units of sites,
The adjoint length is the longest at all couplings, as also found in Ref. [10].
Entanglement. Cutting the infinite chain at one of the six bonds of the unit cell divides it into two half-infinite parts. The reduced density matrix of one part, , has the squared Schmidt values of that bond as eigenvalues, which follow from the bond matrix of the MPS in mixed canonical form. With SU(3) symmetry they come in multiplets: a Schmidt value in irrep is -fold degenerate. For each cell we compute the entanglement entropy and the entanglement energies ; we report averaged over the six cuts, and the largest eigenvalue of over the cuts together with the irrep that carries it.
Trimerization. With on the nearest-neighbor bonds, the trimerization is on the weak bond [10]. In a period-3 state with bond energies , , where
is the sample standard deviation of the three rung energies of the cell, with mean ; for it equals .
Finite-entanglement scaling. At a critical point a finite- MPS has a finite , with [25], and [26]. We extract an effective central charge from the slope of versus over the three most recent states up to and including (two states whose differ by less than 10%, which occur where a run was resumed, are merged); in a gapped phase as saturates.
Extrapolation of the trimerization. Every finite- SU(3)-symmetric state has , also in a gapless phase (see § 2 C). Therefore, the value of at a given bond dimension does not indicate whether the chain is trimerized; the question is whether tends to zero or to a finite value as . A finite- MPS describes the infinite system up to the length scale , which plays the role of a finite system size. In a gapless phase an order parameter that vanishes in the thermodynamic limit then decays as [27, 28], where is the scaling dimension of the corresponding operator [29, 30]. For the trimerization in the SU(3) phase this is the operator , built from the primary field in with , so that . In a gapped phase instead approaches a finite limit once exceeds the correlation length of the ground state. We therefore fit
to the states in the larger- half of each series (at least four), with unless stated otherwise. In the gapless phase this form describes the leading finite- behavior, with up to logarithmic corrections near the Sutherland point (§ 2.4); in the trimerized phase it is an interpolation between the critical behavior at short scales and the finite limit at long scales. Since stays finite, is not a measurement of the order parameter at but a comparative criterion: we compare it with the values obtained in the same way at gapless reference points (, as suggested by field-theory arguments), which measure the bias of the extrapolation and test its dependence on by varying the exponent between and .
Convergence and bond dimensions
Of the 315 states used in this work, 278 reach and 295 reach . States at – and that reached the iteration limit were continued at fixed for up to 1500 further iterations; this changed the energy by less than per site, by at most and by at most 0.01 sites. We reach for –, for , –, – and –, and for – and ; other couplings are taken to –.
Benchmark. At and , , above the exact value. grows as (prediction for : ) and , , , in the windows , , , , approaching from below. Both the reduced exponent and the slow approach of to are expected from the logarithmic corrections generated by the marginally irrelevant coupling of the Sutherland chain [10, 31, 32, 33].
Overview of the phase diagram
Figure 2 summarizes the phase diagram. At small the chain is gapless, as at the Sutherland point; above a lower transition it is trimerized, a threefold-degenerate period-3 valence-bond solid whose order parameter is largest near ; above an upper boundary it is gapless again, and at large the effective central charge follows the crossover from two decoupled chains towards SU(3). The two boundaries are determined in § 2.4 and § 2.5, and the large- regime is discussed in Sec. VI. Here, we describe the ground state in the three regions; the trimerization, the largest Schmidt weight, and the inverse octet correlation length are shown versus for several bond dimensions in Figure 8.
Figure 2. Phase diagram along (linear scale at and logarithmic above). Shaded: gapless SU(3) phase (), trimerized phase, and gapless phase at large , where the SU(3) SU(3) crossover is expected; grey band: –, where the upper boundary lies but is not resolved (, § 2.5). Red: trimerization extrapolated to . Blue: effective central charge at the largest bond dimension; dashed and dotted red lines mark and .
Figure 8. SU(3)-symmetric VUMPS data at bond dimensions matched to the of Ref. [10] (, so ). (a) Trimerization (cf. Fig. 3 of Ref. [10]), (b) largest eigenvalue of the half-chain reduced density matrix (cut between trimers; cf. Fig. 4, top), (c) inverse octet correlation length (cf. Fig. 5), (d) entanglement entropy versus with fits over all (cf. Fig. 7). The grey band marks the range – in which the iDMRG data of Ref. [10] jump ( to ). No jump occurs in the SU(3)-symmetric data.
Energy. The ground-state energy per site (Figure 3) is smooth over the whole range. Since with , the Hellmann–Feynman theorem gives ; numerical derivatives of agree with the measured to (median deviation at ; with states up to near , to ). The energy has a shallow maximum at –, where changes sign; this maximum is unrelated to the transition at .
Figure 3. Ground-state energy per site. Dotted line: exact value at [7]. Inset: from finite differences (points) and the measured energy of the bonds (line), equal by the Hellmann–Feynman theorem; the dashed line marks the maximum of , where changes sign. States at (inset) and the largest (main panel).
Bond energies. Figure 4 shows all bond energies of the six-site cell at . At the bonds are uniform within , a finite- residual of the translation-symmetry breaking discussed in § 2 C. At the trimers are visible directly: two strong bonds () and one strong bond () form a triangle, separated from the next one by weak bonds ( and ). At and the energy resides in the legs ( to ) and the bonds are frustrated (); a period-3 modulation remains in both bond types and decreases with .
Figure 4. Bond energies over two unit cells at . Line width is proportional to ; blue: negative, red: positive. Numbers give the bond energies of the first unit cell; the pattern repeats in the second.
Entanglement spectrum. Resolved by SU(3) irrep (Appendix D), the entanglement spectrum shows the structure of the trimerized phase directly: at the cut, which separates two trimers, is dominated by a single singlet level (), so that the state is close to a product of SU(3) singlets on three sites. The singlet–octet entanglement gap on this cut varies smoothly with and shows no jump at any . The singlet carries the largest Schmidt value on the cut for all .
The lower transition
At a marginal current–current coupling changes sign [10]; the gap opens as , and just above the transition the correlation length exceeds any length accessible to finite-entanglement methods. Indeed, at the states at – have – sites, still growing as –, and for , which cannot be distinguished from a gapless state by or alone.
The extrapolated trimerization resolves the transition once the bond dimension is large enough (Figure 5). With the extrapolation is biased: is negative for all and at the Sutherland point (). The bias comes from the logarithmic corrections generated by the marginal coupling, which the power law does not capture; it decreases with (at : , , for , 2200, 4100), and it is smallest close to , where the marginal coupling vanishes. Using states up to , is zero within for – and positive and increasing from on: , 0.0006(2), 0.0010(2) and 0.0065(7) at , 0.49, 0.50 and 0.55. Here the exponent is not only assumed: fits with a free exponent give 0.63–0.67 at –0.50, consistent with . This gives , in agreement with the level-crossing analysis of ED spectra, [10], which locates the sign change of the marginal coupling directly. Because the order parameter is exponentially small just above , the data cannot exclude a transition slightly below 0.48; we adopt .

Figure 5. Extrapolated trimerization near the lower transition, from states with . Dotted lines: at the gapless Sutherland point for the same , a measure of the bias of the extrapolation. Dashed line: from ED level crossings [10].
A fit of the essential-singularity form expected for this transition does not improve on this (Appendix A). With SU() symmetry the gap opens as , so that ; fitted to , this form gives (statistical error ), well below , and moves to 0.40–0.49 when a constant offset is added to absorb the bias of the extrapolation.
Upper boundary of the trimerized phase
We locate from the bond-dimension dependence of the effective central charge and of the extrapolated trimerization, calibrated on two reference points (Figure 6a). At the Sutherland point –1.98 for all from 2000 to 6000, as for a gapless state with . Deep in the trimerized phase, at , falls to 0.01 and grows only slowly, to 35 sites at . A gapped state whose correlation exceeds shows
Figure 6. Effective central charge versus bond dimension: slope of versus over the three most recent states up to , plotted at . Where the main run of a coupling starts above , the states below its first come from a second run started at ; at the junction both runs give the same state within the difference in . (a) Across the upper boundary, with the gapless () and gapped () references. (b) Large , including , 26 and 32; lines: (two decoupled chains), (SU(3)) and 2 (SU(3)). probes the renormalization-group trajectory at the length scale ; since the crossover lengths grow with , at fixed it increases with , and the values in (b) are not a plateau at (see Sec. VI).
decreasing with , through and below 2; a gapless state in the SU(3) phase approaches , and in the crossover regime discussed in Sec. VI stays above it.
For to 3.5, lies below 2 and keeps decreasing up to the largest bond dimensions: (), 1.41 (2.75, both at ), 1.45 (3.0), 1.56 (3.25) and 1.68 (3.5, all at ). At and 4.0 it is 2.24 and 2.35 (), decreasing slowly. The curves for – have kinks, where drops by up to 0.5 between consecutive points. They do not come from the entropy: the two leading eigenvalues of the octet sector of the transfer matrix are nearly degenerate ( for ) and cross at , 4000 and 5000–6000 for , 3.25 and 3.5, where switches to the faster-growing branch. With the length , which is smooth through the crossings, decreases smoothly with and reaches 1.61, 1.66, 1.78 and 1.91 at , 3.0, 3.25 and 3.5 at the largest , against 1.97 at ; with the values are lower still. Below 2 for every choice of the length, thus indicates a gap up to . For the outcome depends on the length used (2.01 and 2.09 at and 4.0 with , below 2 with ), so that does not locate the boundary there. Quantities that use no correlation length, the growth rate of the entropy and the decay rate of the trimerization, decrease with at every –10, by 15–40% over our range of , and change smoothly with ; only at do they approach zero (0.04 and 0.03 at ).
The extrapolated trimerization (Figure 7) gives the same picture: , 0.022 and 0.019 at , 3.0 and 3.25, 0.006 at both and 3.75, and at , compatible with zero given the bias of the extrapolation (the same fit gives at , in the gapless phase). Varying the number of states in the slope from three to five changes by at most 0.14 (0.27 at ), without moving any value across .
Figure 7. Trimerization extrapolated to , (points, exponent 2/3), the range of for exponents between 1/2 and 1 (band), and at the largest (line); inset: the region near . Open points: couplings with near the upper boundary (, 3.6), not included in the band. Dashed: ; grey band: –5, where the upper boundary lies but is not resolved.
Because Eq. (5) with describes only the gapless SU(3) phase (§ 2), Figure 7 also shows for exponents between 1/2 and 1 (band). Deep in the phase hardly depends on the exponent (0.351–0.354 at ), and fits with a free exponent give 1.0–1.6 for –3.5 and larger values of , so that is the conservative choice there. At the gapless reference points , 5 and 6 the extrapolation gives values between −0.011 and +0.006 over this range of exponents, which sets the resolution of the criterion . For every exponent in the range, at lies above the values at and 6 (e.g. 0.016 against 0.006 and 0.003 for exponent 1), whereas at it lies only slightly above them. Using the singlet correlation length instead of in the extrapolation shifts upward by 0.003–0.007 at all –6, including the gapless reference points, and leaves this comparison unchanged.
We conclude that the trimerized phase extends at least to : lies clearly above the gapless reference values up to (0.02–0.03) and above them for every exponent at , and is below 2 for every choice of the correlation length up to . At , the data behave as at the gapless reference points; where the trimerized phase ends between and 5 is not resolved at the bond dimensions reached here.
Large : the SU(3)→SU(3) crossover
For SU(), : for two decoupled SU(3) chains, 16/5 for SU(3) and 2 for SU(3). Beyond the field theory predicts that the inter-leg coupling first gaps the sector, leaving SU(3), which then flows without a gap to SU(3) [10, 11].
With SU(3) symmetry the effective central charge decreases from the two-chain value towards the SU(3) value (Figure 6b). For all –10 it lies below 16/5 at : 2.35 () and 2.62 (5) at , and 2.89–3.16 for –10 at . At and 5 it keeps decreasing with (at from 2.81 at to 2.62 at ), and at every –10 the values from states with are lower than from states with .
The departure from 16/5 is seen first at the smallest , as expected for the massless flow, whose crossover length grows as the inter-leg coupling decreases.
The values in Figure 6b are therefore set by the accessible bond dimension rather than forming a plateau: measures the central charge at the length scale , and the crossover lengths grow with , the gap length as [10]. At –32, where reaches 86–94 sites at , stays at 3.4–3.6, between 16/5 and 4, and the crossover is not resolved. Reaching itself would require bond dimensions well above even at and 5.
Comparison with previous work
Exact diagonalization
Our bound excludes the ED estimate [10]: at –2.5 all our diagnostics show a gapped, trimerized state, with –71 sites at and –0.05. Since the correlation length near the upper boundary exceeds 70 sites, rings of up to 27 sites used in ED are far from the thermodynamic limit there; moreover, level spectroscopy assumes that the transition is described by SU(3) with a marginal perturbation, which holds only beyond the crossover length of the SU(3) →SU(3) flow (Sec. VI). It is compatible with the position of the jump in the largest- iDMRG data [10], but is obtained independently of it.
iDMRG with Abelian symmetry
Figure 8 shows our data in the same form as the iDMRG data of Ref. [10], at bond dimensions matched to their (both count states on a bond). For the two calculations agree well: the inverse octet correlation length at is 0.016, 0.0095, 0.0055 for , 512, 1024 (Ref. [10]: 0.019, 0.010, 0.006), the trimerization peaks at –0.385 near (0.38), the largest reduced-density-matrix eigenvalue at is ( at ), and the central charge at is (Ref. [10]: ).
In the iDMRG data of Ref. [10], all these quantities jump at a coupling that depends on , from at to at , after which is zero. Our data show no such jump at any up to $8000 [Figure 8(a)–(c) show ; grey band: range of the jumps]. Because an injective SU(3)-symmetric MPS cannot become translation invariant, the absence of a jump in our data is not by itself evidence against a transition; the origin of the jumps is examined below and in Appendix B.
At large , Ref. [10] found – for – with U(1)U(1) symmetry and , close to the SU(3) value. With SU(3) symmetry and we find below for all – and decreasing with at and (Sec. VI).
Origin of the iDMRG jumps
Could the jumps of Ref. [10] be a switch between two different SU(3)-symmetric states? Such states exist: on the cuts the largest Schmidt value can be carried by the singlet (“singlet type”) or by the octet (“octet type”), and in our scans the dominant irrep changes from the singlet to the octet near . We therefore optimized both types at the same and (–, , and ), starting from a state of each type, and compared their energies. Where the iDMRG data jump, the singlet type is clearly the ground state: for it is lower at every , by – per site, far above the convergence error. The two types become nearly degenerate only for , where their energy difference is per site at and 970 and changes sign with . A switch between SU(3)-symmetric states therefore cannot produce the jumps, and the change of the dominant irrep near is a crossing of two nearly degenerate approximations, not a phase transition.
Since the position of the iDMRG jumps moves with , it is not a measurement of a phase boundary; the boundaries determined in § 2.4 and § 2.5 from the -scaling of , and do not rely on it. Finally, we checked directly that the jumps are reproduced when only Abelian color conservation is imposed, using an independent infinite-DMRG code (Appendix B). At its energy and entanglement entropy agree with those of the SU(3)-symmetric state with the same number of states up to (within per site and 0.01), and the entropy drops between and 3.3; at the Abelian state is less entangled ( against 2.167) and lower in energy by per site.
We also test the symmetry from the Schmidt values and the color charges of the Schmidt states: in an SU(3)-symmetric state the Schmidt values form degenerate multiplets, and the summed Schmidt weight at each color weight is invariant under permutations of the three colors. Before the jump, the Abelian states are close in energy but already partially breaking SU(3)-symmetry—the effect decreases with : at the eight Schmidt weights following the dominant singlet, which carry the color charges of an octet, have a relative spread (max−min)/mean of 0.41, 0.14 and 0.02 at , 64 and 128, and at the color-permutation asymmetry of the Schmidt weights is at most 0.022 for . After the jump, at and , the asymmetry is 0.37 and the octet weights spread by twice their mean (Table I): the state breaks SU(3). The same comparison at shows the motion of the jump: at and 3.15 the states lie – below the SU(3)-symmetric ones, with strongly split octet weights, whereas at the same couplings are again close in energy to the SU(3)-symmetric branch. The broken state is not the translation-invariant superposition of three shifted SU(3)-symmetric states either: each component would carry only about states, giving an energy close to that of an SU(3)-symmetric state with , well above the SU(3)-symmetric state with , and an entropy larger by . A continuous symmetry with a non-conserved order parameter cannot be broken spontaneously in the ground state of a one-dimensional chain [34], and the advantage of the broken state is indeed confined to small bond dimension: at its energy at , , lies above that of the SU(3)-symmetric state at , . The jump is thus a crossing, at fixed bond dimension, to an SU(3)-breaking state that is favorable only at small , and its position moves to larger with increasing , as in Ref. [10].
Conclusions
Using uniform matrix product states with exact SU(3) symmetry at bond dimensions up to , we have determined the phase diagram of the SU(3) – Heisenberg chain (Figure 2). A gapless SU(3) phase at small is followed, from , by a threefold-degenerate trimerized phase, whose order parameter and correlation length are largest and shortest, respectively, near (, sites). The trimerized gapless state extends at least to , and at the chain is gapless; where between the two it ends is not resolved at the bond dimensions reached here. This resolves the disagreement between exact diagonalization () and iDMRG () in favor of the larger value, which we obtain from the bond-dimension scaling of the entanglement and of the trimerization rather than from the position of a jump.
The jumps found in iDMRG with U(1)U(1) symmetry are crossings, at fixed bond dimension, to SU(3)-breaking states that are favorable only at small bond dimension. An injective SU(3)-symmetric MPS cannot be invariant under one-site translation (a translation-invariant one exists only as a non-clustering superposition of three translates), so finite- SU(3)-symmetric states always carry some trimerization and show no such jumps. Phase boundaries in SU() chains with the fundamental representation on each site should therefore be located by finite-entanglement scaling, as done here, or by spectroscopy that tracks the sign change of the marginal coupling directly.
At large the effective central charge lies below the SU(3) value for all –10 and decreases with bond dimension where the crossover length is shortest, as expected for the massless flow to SU(3).
Three questions remain open. Reaching the SU(3) value requires correlation lengths beyond the crossover length of the SU(3)SU(3) flow, i.e. bond dimensions well above those used here, the more so at larger . The position of above 3.5 and a possible transition slightly below are limited by the exponentially slow opening of the gap at both boundaries; a level-spectroscopy analysis of the transfer matrix or of excitations resolved by irrep would determine both boundaries independently of finite-entanglement scaling.
Acknowledgments
The authors acknowledge support from the Simons Foundation in the framework of the call “Centers of Excellence in Ukraine”, grant No. SFI-PD-COE-00032628.
Essential-singularity fits
Figure 9 shows fits of on the window . Here is extrapolated from all available states of each coupling with : up to at –, – at and –, and at . The window is fitted with the essential-singularity form expected for the lower transition. At , a marginal current–current coupling changes sign [10]. With SU() symmetry, it is the only coupling that flows , so that strong coupling is reached after and . Since ,
Figure 9. Fit of the SU() essential-singularity form, Eq. (6), to the extrapolated order parameter, using the singlet (left) and octet (right) correlation length in the extrapolation. The vertical line marks . The plotted quantity is .
for . With extrapolated using the singlet or the octet correlation length, the fit gives and ( and ), with and ; on the narrower windows and it gives –, with up to 58. Adding a constant offset as a free parameter (to absorb the bias discussed in § 2.4) moves to –. The location of the transition cannot be obtained in this way at reachable bond dimensions.
Abelian iDMRG check
The Abelian calculations use an independent infinite-DMRG code (ITensors [35]) with the same six-site cell and Hamiltonian convention as the VUMPS calculations, and final bond dimensions and . The code conserves the numbers of particles of two of the three colors modulo 6, i.e. a subgroup of the color symmetry used in Ref. [10], which keeps the number of charge sectors on the bonds of the infinite chain bounded; like Ref. [10], it does not impose SU(3). Its local bond
| 2.0 | 32 | — | 0.009 | 0.41 |
| 2.0 | 64 | 0.015 | 0.14 | |
| 2.0 | 128 | 0.002 | 0.02 | |
| 3.0 | 32 | — | 0.286 | 2.43 |
| 3.0 | 64 | 0.034 | 0.80 | |
| 3.0 | 128 | 0.018 | 0.48 | |
| 3.15 | 32 | — | 0.285 | 2.43 |
| 3.15 | 64 | 0.039 | 0.87 | |
| 3.15 | 128 | 0.022 | 0.60 | |
| 3.5 | 32 | — | 0.281 | 2.44 |
| 3.5 | 64 | 0.047 | 0.99 | |
| 3.5 | 128 | 0.371 | 1.98 |
Table 1. Abelian iDMRG states compared with SU(3)-symmetric VUMPS states with the same number of states: energy difference per site (no SU(3)-symmetric state with , so none at ), color-permutation asymmetry of the Schmidt weights, and spread of the octet weights. Both and vanish for an SU(3)-symmetric state. The jump at lies between and 3.3.
energies are evaluated on a finite block with approximate environments and do not show the period-3 structure of the infinite state; we therefore compare only the energy per site and the entanglement entropy, which the code reports for one cut. The SU(3)-symmetric reference states for this comparison were computed at small bond dimension () by two-site growth and VUMPS, as in the main calculations; their energies and least-entangled-cut entropies are interpolated linearly in to . At the Abelian and the SU(3)-symmetric energies agree within per site for –, and the entropies within 0.01. The entropy of the Abelian state rises smoothly up to () and drops to 2.077 at and 2.074 at 3.5, where the SU(3)-symmetric state has $S = 2.167 (Figure 10).
Figure 10. (a) Entanglement entropy versus (, so ): SU(3)-symmetric VUMPS at (least-entangled cut) and Abelian iDMRG (color numbers conserved modulo 6) at and 128. (b) Energy difference per site between the Abelian state at and the SU(3)-symmetric state with the nearest bond dimension (–141).
Figure 11. Entanglement spectra resolved by SU(3) irrep on the two inequivalent cuts (triality ; the cut is the conjugate of ), at the Sutherland point (), in the trimerized phase (), near the upper boundary () and at large (). Each level of irrep is -fold degenerate; denotes the irrep with Dynkin labels .
To test the symmetry of the Abelian states we record, at the end of each stage, the Schmidt values of the central cut together with the color charges mod 6 of the Schmidt states. Taking the charges relative to the largest Schmidt state, with , we form the summed weight of the Schmidt states with color weight and the asymmetry
which vanishes for an SU(3)-symmetric state. The test is necessary but not sufficient, since a splitting inside a multiplet that keeps the color-permutation symmetry leaves ; we therefore also record the relative spread of the eight weights , which in an SU(3)-symmetric state form the octet following the singlet. Table I lists both together with the energy difference. Before the jump both measures decrease with ; at and the state is strongly SU(3)-breaking. Repeated runs of the code at reach different broken states, with energies differing by up to $5.5 \times 10^{-3}\chi = 128 results are reproduced exactly; the table uses one set of runs throughout.
No one-site translation invariance of injective SU(3)-symmetric MPS
Figure 12. Singlet–octet entanglement gap on the cut, at bond dimensions matched to Ref. [10] and at the largest . Dashed line: ; grey band: –5, where the upper boundary lies (§ 2.5).
Consider a uniform MPS with tensors , , acting on a virtual space of dimension , and its transfer map . The MPS is injective (normal) if, after normalization, is the only eigenvalue of of modulus one, it is non-degenerate, and its eigenvector is positive definite; equivalently, connected correlations decay exponentially and the state is the unique ground state of a local parent Hamiltonian. We show that such an MPS cannot be invariant under SU(3) acting in the fundamental representation on every site.
(i) If an injective MPS is invariant under for all , there are a unitary representation on , defined up to phases, and a one-dimensional representation such that [22, 23, 24, 36].
(ii) SU(3) has no non-trivial one-dimensional representations, so ; it is simply connected, so can be chosen as a linear representation and decomposes into irreps. By Schur’s lemma the center element , , acts on an irrep of triality as , so that , where collects the irreps of triality .
(iii) In the fundamental representation . Hence : maps to .
(iv) Using , . Since is invertible and is positive definite, ; thus is a second eigenvalue of of modulus one, which contradicts injectivity.
The same argument applied to the tensor of an -site cell, on which the center acts as , shows that an injective SU(3)-symmetric MPS requires ; the six-site cell used here satisfies this. If the injectivity assumption is dropped, a one-site invariant SU(3)-symmetric MPS exists: with cyclic between the blocks, for which has the eigenvalues , and on the unit circle. In the thermodynamic limit this state is the equal-weight superposition of the three translates of a period-3 injective state; its correlations do not cluster. The argument is the SU(3) counterpart of the one for SU(2) with half-integer spin, where the center plays the role of [22, 23, 24].
Entanglement spectra
Figure 11 shows the entanglement spectra, with the eigenvalues of the half-chain reduced density matrix, resolved by SU(3) irrep. At the cut is dominated by a single singlet level (), separated by from the first octet, and the neighboring cuts by a single or , as for a product of SU(3) singlets on three sites. At , and the spectra are dense, with many irreps at low entanglement energy. Figure 12 shows the gap between the lowest octet and the lowest singlet level on the cut (cf. Figure 4, bottom, of Ref. [10]). It peaks at , where it is independent of , and decreases slowly with outside the trimerized phase. It shows no jump at any .
Code availability
The code used to obtain and analyze the data (based on MPSKit and TensorKit [16, 17]) will be made publicly available upon publication.
References
References
- [1]A. V. Gorshkov, M. Hermele, V. Gurarie, C. Xu, P. S. Julienne, J. Ye, P. Zoller, E. Demler, M. D. Lukin, and A. M. Rey, Two-orbital SU(N) magnetism with ultracold alkaline-earth atoms, Nat. Phys. 6, 289 (2010).DOI
- [2]M. A. Cazalilla and A. M. Rey, Ultracold Fermi gases with emergent SU(N) symmetry, Rep. Prog. Phys. 77, 124401 (2014).DOI
- [3]S. Taie, R. Yamazaki, S. Sugawa, and Y. Takahashi, An SU(6) Mott insulator of an atomic Fermi gas realized by large-spin Pomeranchuk cooling, Nat. Phys. 8, 825 (2012).DOI
- [4]G. Pagano, M. Mancini, G. Cappellini, P. Lombardi, F. Schäfer, H. Hu, X.-J. Liu, J. C. Catanio, C. Sias, M. Inguscio, and L. Fallani, A one-dimensional liquid of fermions with tunable spin, Nat. Phys. 10, 198 (2014).
- [5]X. Zhang, M. Bishof, S. L. Bromley, C. V. Kraus, M. S. Safronova, P. Zoller, A. M. Rey, and J. Ye, Spectroscopic observation of SU(N)-symmetric interactions in Sr orbital magnetism, Science 345, 1467 (2014).DOI
- [6]C. Hofrichter, L. Riegger, F. Scazza, M. Höfer, D. R. Fernandes, I. Bloch, and S. Fölling, Direct probing of the Mott crossover in the SU(N) Fermi-Hubbard model, Phys. Rev. X 6, 021030 (2016).DOI
- [7]B. Sutherland, Model for a multicomponent quantum system, Phys. Rev. B 12, 3795 (1975).
- [8]P. Corboz, A. M. Läuchli, K. Totsuka, and H. Tsunetsugu, Spontaneous dimerization in a bilinear-biquadratic S=1 zig-zag chain, Phys. Rev. B 76, 220404 (2007).DOI
- [9]S. Rachel, R. Thomale, M. Führinger, P. Schmitteckert, and M. Greiter, Spinon confinement and the Haldane gap in SU(n) spin chains, Phys. Rev. B 80, 180420(R) (2009), arXiv:0904.3882.DOI
- [10]L. Herviou, S. Capponi, and P. Lecheminant, Even-odd effects in the J1-J2 SU(N) Heisenberg spin chain, Phys. Rev. B 107, 205135 (2023), arXiv:2302.14090.DOI
- [11]P. Lecheminant, Massless renormalization group flow in SU(N)k perturbed conformal field theory, Nucl. Phys. B 901, 510 (2015), arXiv:1509.01680.DOI
- [12]P. Nataf and F. Mila, Exact diagonalization of Heisenberg SU(N) chains in the fully symmetric and antisymmetric representations, Phys. Rev. B 93, 155134 (2016), 1601.00922.DOI
- [13]P. Nataf, S. Gozel, and F. Mila, Edge states and universality class of the critical two-box symmetric SU(3) chain, Phys. Rev. B 104, L180411 (2021), arXiv:2107.09306.DOI
- [14]V. Zauner-Stauber, L. Vanderstraeten, M. T. Fishman, F. Verstraete, and J. Haegeman, Variational optimization algorithms for uniform matrix product states, Phys. Rev. B 97, 045145 (2018), 1701.07035.
- [15]L. Vanderstraeten, J. Haegeman, and F. Verstraete, Tangent-space methods for uniform matrix product states, SciPost Phys. Lect. Notes , 7 (2019), 1810.07006.DOI
- [16]L. Devos, M. Van Damme, and J. Haegeman, MPSKit.jl (2026), version 0.13.13.
- [17]L. Devos and J. Haegeman, TensorKit.jl: A Julia package for large-scale tensor computations, with a hint of category theory (2025), arXiv:2508.10076.arxiv.org/abs/2508.10076
- [18]SUNRepresentations.jl, https://github.com/QuantumKitHub/SUNRepresentations.jl (2026), accessed October 2026.
- [19]A. Alex, M. Kalus, A. Huckleberry, and J. von Delft, A numerical algorithm for the explicit calculation of SU(N) and SL(N,C) Clebsch-Gordan coefficients, J. Math. Phys. 52, 023507 (2011), arXiv:1009.0437.DOI
- [20]I. P. McCulloch and M. Gulácsi, The non-Abelian density matrix renormalization group algorithm, Europhys. Lett. 57, 852 (2002), cond-mat/0012319.
- [21]A. Weichselbaum, Non-abelian symmetries in tensor networks: A quantum symmetry space approach, Ann. Phys. 327, 2972 (2012), 1202.5664.DOI
- [22]E. Lieb, T. Schultz, and D. Mattis, Two soluble models of an antiferromagnetic chain, Ann. Phys. 16, 407 (1961).
- [23]I. Affleck and E. H. Lieb, A proof of part of Haldane’s conjecture on spin chains, Lett. Math. Phys. 12, 57 (1986).
- [24]M. Sanz, M. M. Wolf, D. Pérez-García, and J. I. Cirac, Matrix product states: Symmetries and two-body Hamiltonians, Phys. Rev. A 79, 042308 (2009), arXiv:0901.2223.DOI
- [25]F. Pollmann, S. Mukerjee, A. M. Turner, and J. E. Moore, Theory of finite-entanglement scaling at one-dimensional quantum critical points, Phys. Rev. Lett. 102, 255701 (2009).pubmed.ncbi.nlm.nih.gov/19659096
- [26]P. Calabrese and J. Cardy, Entanglement entropy and quantum field theory, J. Stat. Mech. , P06002 (2004).
- [27]L. Tagliacozzo, T. R. de Oliveira, S. Iblisdir, and J. I. Latorre, Scaling of entanglement support for matrix product states, Phys. Rev. B 78, 024410 (2008).DOI
- [28]B. Pirvu, G. Vidal, F. Verstraete, and L. Tagliacozzo, Matrix product states for critical spin chains: Finite-size versus finite-entanglement scaling, Phys. Rev. B 86, 075117 (2012).DOI
- [29]I. Affleck, Exact critical exponents for quantum spin chains, non-linear σ-models at θ = π and the quantum Hall effect, Nucl. Phys. B 265, 409 (1986).DOI
- [30]I. Affleck, Critical behaviour of SU(n) quantum chains and topological non-linear σ-models, Nucl. Phys. B 305, 582 (1988).DOI
- [31]I. Affleck, D. Gepner, H. J. Schulz, and T. Ziman, Critical behaviour of spin-s Heisenberg antiferromagnetic chains: analytic and numerical results, J. Phys. A: Math. Gen. 22, 511 (1989).
- [32]S. Eggert, Numerical evidence for multiplicative logarithmic corrections from marginal operators, Phys. Rev. B 54, R9612 (1996), arXiv:cond-mat/9602026.DOI
- [33]C. Itoi and M.-H. Kato, Extended massless phase and the Haldane phase in a spin-1 isotropic antiferromagnetic chain, Phys. Rev. B 55, 8295 (1997), arXiv:cond-mat/9605105.DOI
- [34]S. Coleman, There are no Goldstone bosons in two dimensions, Commun. Math. Phys. 31, 259 (1973).DOI
- [35]M. Fishman, S. R. White, and E. M. Stoudenmire, The ITensor software library for tensor network calculations, SciPost Phys. Codebases , 4 (2022).DOI
- [36]D. Pérez-García, M. M. Wolf, M. Sanz, F. Verstraete, and J. I. Cirac, String order and symmetries in quantum spin lattices, Phys. Rev. Lett. 100, 167202 (2008), 0802.0447.









