Arxiv:1906.06553V2 [Cond-Mat.Str-El] 19 Aug 2019 T Al

Home , Universality class

arXiv:1906.06553v2 [cond-mat.str-el] 19 Aug 2019 t al. et. Psaels tde ofrdsiete perfre- appear Remarkably, they despite systems. far many-body quantum so in studied quently less are QPTs rtclpitipisa mretsmer.The symmetry. emergent an implies the point and phases critical (SSB) breaking symmetry spontaneously interac- by fine-tuned a strength detected upon tion order be first of not or continuous ther can parameters it order thus local between phases, place take disordered may two QPT topological The paradigm. (DQCP) such point exceptions cal QPT some witnessed topological have as the- years (LGW) recent Landau-Ginzburg-Wilson ory, well rele- the are by other QPTs of tran- continuous described the and traditional instead, While across parameter is, point. discontinuity sition order other display the observables The vant when order gap. first energy transi- length the vanishing correlation at diverging and a ground continuously by the accompanied changes point, when system tion continuous the is classi- of One be state types. are general two temperature, in into could zero fied and at phenomena occurr omnipresent which (QPTs), sitions Q odsld(B)phase (VBS) valence a solid and Néel weakly bond phase a (or between continuous transition a order) for first evidences provide studies cal a n infiatpolm nmdr odne mat- condensed modern fundamen- in physics tran- most problems ter the the significant of as and one well tal remains as them phases among different sitions into order systems ncnrs otecniuu Ps h first-order the QPTs, continuous the to contrast In Introduction model 7 , 8 9 nrni upCaatro h is-re unu hs T Phase Quantum First-Order the of Character Jump Intrinsic hrb otnosQTocr ewe two between occurs QPT continuous a whereby ssc neapeweeetnienumeri- extensive where example an such is 1 – nrdcdt ug hte P so rtodro o.This not. or order first of is QPT spin-1 a solvable whether exactly judge to introduced hr itntjmsof jumps distinct where rps odrvra ehdweeaquantity a where no method are ones reversal continuous bond the a while propose operators relevant of jumps oncigtoodrdpae naspin-1 a in phases anal ordered one-dimensional two proposed connecting recently a study we Finally, rniinsice rmbigcniuu oehbtn fi a exhibiting to ( continuous cross-coupled being a from of switches QPT transition topological the study to rniinblnst h asinuieslt ls with class po universality transition the Gaussian crossing the when to smooth belongs is transition DBS the because QPT 3 13 npriua,teqatmpaetran- phase quantum the particular, In . –Udrtnighwsrnl correlated strongly how Understanding .– efidta h rtodrqatmpaetastos(QPTs) transitions phase quantum first-order the that find We – 16 3 4 h QPwspooe ySenthil by proposed was DQCP The . e aoaoyo rica tutrsadQatmControl Quantum and Structures Artificial of Laboratory Key olbrtv noainCne o dacdMicrostruct Advanced for Center Innovation Collaborative 7 2 – colo hsclSineadTechnology and Science Physical of School 4 10 – 1 antcMtraso h o,LnhuUiest,Lanzhou University, Lanzhou MoE, the of Materials Magnetic 6 11 eateto hsc,Rni nvriyo hn,Beijing China, of University Renmin Physics, of Department htbyn h cp fLGW of scope the beyond that 5 17 , n eofie unu criti- quantum deconfined and 12 ejn opttoa cec eerhCne,Biig10 Beijing Center, Research Science Computational Beijing – 19 htsmr,i ol eei- be could it more, What’s . . colo hsc n srnm,TugDoLeInstitute, Lee Tsung-Dao Astronomy, and Physics of School in Luo, Qiang hnhiJa ogUiest,Saga 020 China 200240, Shanghai University, Tong Jiao Shanghai / 2 D XXZ pera h rtodrtasto onsfrbt ae.W cases. both for points transition first-order the at appear 1, esnegcanadaqatmIigcanwt longitudina with chain Ising quantum a and chain Heisenberg ∗ ieZhao, Jize Dtd uut2,2019) 20, August (Dated: / hi.W ueottepsiiiyo ekyfirst-order weakly of possibility the out rule We chain. 2 J - 2, † J n iou Wang Xiaoqun and a bevdeprmnal nadie ici quantum circuit system breakdown driven electrodynamics a photon-blockade in a experimentally observed called was QPT first-order a aebe icse ni recently QPTs until first-order discussed the been of have points information transition the quantum near from sciences borrowed probes various and ia motnet eieaporaetosfraproper features. a dominating for their tools of appropriate characterization devise to importance vital where epc to respect ue hnbt fte ol esmlaeul diag- of simultaneously independent be are could eigenfunctions them and of onalized both then mute, en httesetaof spectra the that means ua hneo rudsaewv ucin eas of struc- Because the energy of reflects function. continuity wave also the ground-state jump of The change QPT. tural first-order a of odsrnt DS sitoue omgiytejump as the defined magnify is to It introduced behaviors. is (DBS) strength bond loepce oeiiaetesnuaiy odtc the detect To singularity. point the transition eliminate to expected also λ ol aylnal with linearly vary could ee-rsigpoint level-crossing a e site per loocradw oetedsuso otesupplemen- the to (SM) discussion could material the QPT move tal continuous we that and aware occur be also should It sites. tice λ × & D hl h oa rudsaeenergy ground-state total the while , pnlde hr h Haldane–rung-singlet the where ladder spin ) xiisnnnltct.Here, nonanalyticity. exhibits e scnie aitna fteform the of Hamiltonian a consider us Let h ieec fbn tegh(B) is (DBS), strength bond of difference the , e aoaoyfrMgeimand Magnetism for Laboratory Key h eta charge central the g fdcnndqatmciia point critical quantum deconfined of ogy λ s-re hrce at character rst-order e g .Bsdo uha bevto,we observation, an such on Based t. sadiigprmtr If parameter. driving a is ( λ λ rs ajn 103 China 210093, Nanjing ures, n.Mroe,w ffimta such that affirm we Moreover, int. upof jump a , (1 = ) ehdi rtyapidt an to applied firstly is method r hrceie yintrinsic by characterized are λ Mnsr fEducation), of (Ministry t 26 quantity a , D /L H aigtedrvtv of derivative the Taking . ,4 5, 4, 3, = 04 China 0084, 300 China 730000, ( 082 China 100872, ) λ λ E e = ) c t 20 g 0 e 1. = ( λ ‡ hr h rudsaeenergy ground-state the where e h nt-iesaigo gap of scaling finite-size The . I − λ g osqety hr a be can there Consequently, . H = ) ( H at sgn( λ 0 J 0 ,asmlrjm for jump similar a ), D × λ + and e ,I t λ hnueit use then e ubdtedffrneof difference the dubbed 0 λ L 21 ilapa,indicating appear, will t ransitions ≃ H + ) – e stenme flat- of number the is 25 H I I λe 0.30(2). , hrfr,i sof is it Therefore, . I H field l I E r reeatof irrelevant are 0 safnto of function a as g and ( λ 1 = ) e g H ( λ λ I hH This . with ) com- e ( 0 λ (1) (2) ) is i 2 where the minus sign reflects the spirit of the bond re- years later37. According to Eq. (2), the DBS is defined versal method. Nevertheless, in most systems the two as = Sz Sz + 2 Sx Sx . DL h L/2 L/2+1i h L/2 L/2+1i terms 0 and I do not commute, resulting in a cum- At ∆ = 1 the Hamiltonian Eq. (3) posses a hidden H H − bersome expression of eg(λ) versus λ. However, the main FM SU(2) symmetry, resulting in a ground-state mani- spirit remains unchanged in that the jump character of fold of degenerate SU(2) multiplet corresponding to the D faithfully inherits the discontinuity of first-order QPTs. largest total spin. The model is not conformal invari- In what follows we will firstly illustrate the bond re- ant and dramatic changes of its entanglement behaviors versal method in two different but thoroughly studied occur38–41. In Fig. 1(a) we show the ground-state energy one-dimensional (1D) spin models that possess first-order eg around the transition point ∆ = 1. A vivid cusp of QPTs: (i) a celebrated spin-1/2 Heisenberg XXZ chain the energy curve could be spotted at− ∆ = 1, while it for which all the energy as well as the DBS can be cal- turns to be a jump of in Fig. 1(b). This− gives clearly D L culated analytically and (ii) a quantum Ising chain with a first glance of the jumpD character of in the first-order both longitudinal and transverse fields which does not QPT26. D host exact solution but the transition line is well-known. Having established the cornerstone of our method, we -0.2 0.3 then apply it to (iii) a topological QPT of a cross-coupled spin ladder and (iv) a recently proposed spin-1/2 chain 0.1 with DQCP, both of which are beyond the scope of the -0.3 conventional LGW paradigm. All the models are studied by the density-matrix renormalization group (DMRG) -0.1 method27–30, which is a powerful tool for dealing with quantum-mechanical problems in 1D systems. We uti- -0.4 lize the periodic boundary condition (PBC) for the first -0.3 two cases to have a better comparison with analytical results. For the latter cases, however, we turn to the -0.5 -0.5 open boundary condition (OBC) which is beneficial to -2 -1 0 -2 -1 0 large-scale numerical calculations. XXZ chain.– The 1D spin-1/2 Heisenberg XXZ chain has long served as the workhorse for the study of quan- FIG. 1: (a) The ground-state energy eg for L = 64 (red rhom- tum magnetism31. Its Hamiltonian is given by bus), 128 (blue circle), and TDL (black line). (b) The same setup as (a) for DBS DL. L 1 = S+S− + S−S+ + ∆SzSz (3) H 2 i i+1 i i+1 i i+1 Quantum Ising chain with longitudinal field.– The 1D i=1 X quantum Ising chain is integrable and its exact solution 42 ± x y was presented by Pfeuty in 1970 . It owns a continuous where Si = Si + iSi is the raising/lowering operator at QPT of the Ising universality class separating a FM and site i and ∆ is the anisotropic parameter. In particular, a paramagnetic phase at the critical value of transverse in the region 1 < ∆ 1 the ground state is a Lut- 43 − ≤ field hz( 0) . What’s more, an emergent E8 symmetry tinger liquid (LL) with a gapless excitation spectrum. was experimentally≥ verified around the critical point44. The ground-state energy per site eg in the thermody- By introducing a longitudinal field h the model is no namic limit (TDL) L can be calculated as32–34 x → ∞ longer integrable except for a specially fine-tuned weak 45 ∞ longitudinal field . The total Hamiltonian is thus given LL ∆ sin πυ tanh υx eg (∆) = 1 dx (4) by 4 − π 0 − tanh x Z L with ∆ = cos πυ. Beyond the critical region it presents = σxσx + h σz + h σx (5) H − i i+1 z i x i a long-range ferromagnetic (FM) or antiferromagnetic i=1 (AFM) order, exhibiting in correspondence to the FM X whereσ ˆ = (σx, σy, σz) are the Pauli matrices. In the FM point ∆ = 1 a first-order QPT, and a continuous one 24,46,47 belonging to− the Kosterlitz-Thouless (KT) universality phase (hz < 1) of the magnetic phase diagram , a class at the AFM point ∆ = 1. In the FM phase first-order QPT with a discontinuity of the magnetization, M = 1 σx , takes place at h = 0. Specifi- (∆ < 1), the spins are parallel along the z direction, x L ih i i x − FM cally, let hz = 1/2 we have the exact ground-state en- resulting in eg = ∆/4. The spin-spin correlation func- P ergy e = 3 E 6√2/3 1.0635 where E( ) is tions could be obtained by Hellmann-Feynman theorem. g,0 − 4π ≈ − · In the LL phase, however, no simple expressions for the the complete elliptic integral of second kind42. As can p correlation functions are available except for some ratio- be seen from Fig. 2(a), there is a pinnacle at hx = 0 in nal υ-values35. Historically, the explicit expressions of the energy curve, and the symmetric feature is a rem- 26 Zx the correlation functions were first given by Jimbo and iniscence of 2 symmetry. The DBS is defined as L 36 x x x z D Miwa in 1996 , and then simplified by Kato et al. several = σ σ σ + hzσ , which is merely a L/2 − L/2 L/2+1 L/2 3 shift of M by e currently. In this occasion plays firmly established that such a first-order QPT remains x g,0 DL the role of order parameter Mx and there is no wonder present for deviations away from the dual line as large as that it exhibits a jump at hx = 0 (see Fig. 2(b)). In J× = 0.6, unanimous conclusion has not been drawn on general, since DBS has an ambiguous relation with order whether the first-order QPT could extend to all the locus parameter, it’s thus well founded to regard this jump as of the phase boundary or just end at a nonzero inflection 26 a signal for a first-order QPT . point J×,I . At weak interchain couplings, an early ana- lytic result predicted that the transition is always of first -1.0 0.5 order57, and later a numerical calculation of the same group yields to the conclusion58. Meanwhile, in the work of Wang13, it is found that the first-order QPT is dis- -1.2 -0.5 missed at J×,I = 0.287, and a continuous QPT down to the vanishing interchain couplings takes over afterward. It’s worth mentioning that the fact that a continuous QPT occurs at J× = 0.2 is checked by tensor network -1.4 -1.5 approach66 and quantum Monte Carlo method50. In view of the ambiguity, it is our purpose to determine the inflection point J×,t accurately by bond reversal method. -1.6 -2.5 -0.4 0.0 0.4 -0.4 0.0 0.4

FIG. 2: (a) The ground-state energy eg for L = 64 (red rhom- 0.6 bus), L = 96 (geren square), and 128 (blue circle). (b) The same setup as (a) for DBS DL. 0.3

Cross-coupled spin ladder.– The role of frustra- 0.0 tion in quasi-1D magnetic materials has attracted nu- 0.6 31 merous attention ever since the discovery of high- -0.3 0.5 temperature superconductivity in 1980s48. The cross- 0.4 13,49,50 0.2 0.4 0.3 coupled spin ladder , in particular, is one of the 0.6 0.8 most outstanding models which is not only of theoreti- 1.0 1.2 0.2 cal importance52–54 but also experimentally accessible51. 0.15 The Hamiltonian of the model reads as follows: .10 L L .05 =J S S + J S S H k i,α · i+1,α ⊥ i,1 · i,2 i=1 α=1,2 i=1 X X X 0.00 L 0.2 0.3 0.5 0.6 J S S + S S , (6) × i,1 · i+1,2 i,2 · i+1,1 i=1 FIG. 3: (a) The DBS DL for L = 64 (cyan rhombus), L X = 128 (magenta square), and L = 192 (blue triangular) of S where i,α denotes a spin-1/2 operator at site i of the different J×’s. The solid lines are guided for the eyes. In α-th leg. Jk(= 1) and J⊥ are the NN interactions along the bottom projection plane, the thick black line and the thin the leg and rung directions, respectively. J× > 0 is the black line are the continuous and first-order phase boundaries, antiferromagnetic cross-coupled interaction. respectively. The pentagram (⋆) marks the inflection point. Whereas a continuous QPT with a central charge c =2 (b) The Haldane gap ∆ along the phase boundary. 55 occurs at J⊥ = 0 in the absence of J× , contentious results with a decade disputing exist for nonzero J×. On the one hand, a columnar dimerized phase was predicted During each calculation we shall fix J× and vary J⊥ of between the Haldane phase and the rung-singlet phase in Eq. (6). We thus define the DBS as L = L L where = Sz Sz + Sz DSz R −and L a narrow parameter region at weak cross-coupled inter- L L/2,1 L/2,2 L/2+1,1 L/2+1,2 R z z z z action J 59. Though some clues for the dimerized phase L = S + S S + S in a pla- × L L/2,1 L/2,2 L/2+1,1 L/2+1,2 appear at finite-size case60,61, people now generally be- quette in the spirit of Eq. (2). In Fig. 3(a) we show 62–66 lieve that there is no such a phase actually . On the the curvatures of DBS for different J×’s from 0.2 to 0.6. other hand, when J× = 1, the Hamiltonian of Eq. (6) Here, we keep as large as 2000 states typically in our 49 undergoes a first-order QPT at J⊥,t = 1.401484 . Due DMRG calculation and extend to 3000 states when nec- 56 to the dual symmetry of Eq. (6) , we shall just con- essary. For J× = 0.5 and 0.6, there is a jump of DBS centrate on the case where J× is below the dual line in each case, indicating that a first-order QPT occurs. J× = 1. Though various numerical calculations have For other cases that are smaller than J× = 0.4, how- 4 ever, the curves are rather smooth and no conspicuous plies that the transition is indeed not a first-order one. jumps are encountered. This is a strong evidence that the transitions here are not of first order but continuous -0.25 0.12 with a central charge c =226. Whereas the curvatures of DBS for J× = 0.4 seem to be smooth, a jump which is a signal for first-order QPT appears for large enough system size26. We also calculate the energy gap of Haldane -0.30 0.08 z z phase, i.e., ∆L = Eg(Stot = 2) Eg(Stot = 0), and the results are shown in Fig. 3(b). It− could be found that the gap is infinitesimal within our numerical precision when -0.35 0.04 J× . 0.30, and it opens exponentially afterward. After a series of careful calculations we thus conclude that the inflection point J×,I is a finite value of 0.30(2). Spin-1/2 chain with DQCP.– Whereas the DQCP was -0.40 0.00 originally proposed in two-dimensional systems7,8, the 1.2 1.4 1.6 1.8 1.2 1.4 1.6 1.8 1D analogy of DQCP was constructed quite recently67 and it has been studied by several parallel works on frustrated spin-1/2 chains with discrete symmetries68–70. For FIG. 4: (a) The ground-state energy eg for L = 64 (cyan rhombus), L = 128 (magenta square), L = 192 (blue triangu- concrete, we consider the following anisotropic model, lar), and 256 (red circle). (b) The same setup as (a) for DBS D L L. = J SυSυ + K SυSυ , (7) H − υ i i+1 υ i i+2 i=1 υ=x,z Because of the OBC utilized in our simulations, the X X VBS phase only has a unique ground state while the where Jυ,Kυ > 0 so that the NN interactions are fer- zFM phase still has two-fold degeneracy. We define the romagnetic while the 2nd-NN interactions are antiferro- energy gaps ∆ = E E , as the total energy dif- 1,2 1,2 − g magnetic. We shall treat the NN interaction Jx = 1 ference between the first/second excited states E1,2 and and fix the 2nd-NN interaction K = Kx/z = 1/2 so the ground state Eg. In Fig. 5(a) we show energy gaps that the only adjustable parameter is Jz(> 0). When ∆1,2 versus Jz. With the increasing of Jz, ∆1 decreases Jz is not very large, the ground state of Eq. (7) could all the way and vanishes rapidly when crossing the crit- m be continuously connected to that of the Majumdar- ical point. For ∆2, However, there is a minimum ∆ 71 2,L Ghosh point where Jz = 1. This phase is the well- at each length L around the critical point. As shown in known dimerized VBS phase which breaks translational m the inset, the ∆2,L’s follow a linear scaling versus 1/L symmetry. On the contrary, in the regime where Jz is and the gap at TDL is 0.0000(4), indicating the closure dominant the spins align parallelly along their z direc- of energy gap at the critical point. Because of the linear Zz tions, resulting in a zFM phase with breaking 2 sym- scaling ansatz, the critical point is conformal invariant72. metry. In the original work of Jiang et. al.67, the VBS– The von Neumann entropy (vNE) L is calculated by zFM transition was argued to be continuous, a transi- the minimal entangled ground stateS and the final result tion of which is at odds with the LGW theory where a is shown in Fig. 5(b). A hump appears near the critical direct transition between two states breaking irrelevant point, and this is another evidence for a continuous QPT. symmetries should be of first order. The critical point We fit the maxima of vNE L as a function of length L, Jz,c is called DQCP in analogy with its two-dimensional m = c ln 2L + c′, where Sc is the central charge and c′ SL 6 π counterpart, and a continuous O(2) O(2) symmetry is a nonuniversal constant73. We find that c 1.02(5) at 67 × emerges . The model Eq. (7) has been studied by ma- the critical point. ≃ trix product state (MPS) which works directly in the 68,69 We now calculate the critical point Jz,c and critical TDL in two independent calculations . Both order exponents of the order parameters. The VBS phase parameters of the SSB phases have a tiny but finite jump is characterized by the difference of the adjacent bond around the critical point. Notwithstanding, such discon- VBS S S S S strength, i.e., ML = i i+1 i−1 i . The tinuity is argued to be an artifact of MPS method. In zFM phase has a nonzero|h local· momenti − h at each· sitei| and fact, the weakly first-order phase transition is hardly dis- zFM z 17–19 thus ML = Si . In practice, we could set i = L/2 tinguishable from a continuous one , and thus metic- to minimize the|h finite-sizei| effect. Also, when calculating ulous calculations should be carried out to check the type zFM the ML , a finite pinning field of order 1 is added at of the transition. We therefore resort to DMRG method the boundaries of the open chain so as to select a deter- where up to 2000 states are kept to revisit this problem. minate ground state. Theoretically, the order parameter To begin with, we calculate the ground-state energy M versus J with the length L follows74 and the energy curves shown in Fig. 4(a) is rather L z smooth. The DBS = Sz Sz Sx Sx −β/ν 1/ν DL L/2 L/2+1 − L/2 L/2+1 ML(Jz) L fM Jz Jz,c L , (8) (see Fig. 4(b)) is continuous likewise when tuning J and ≃ | − | z no overt jump could be observed in the curves. This im- where the critical exponent ν describes the divergence of 5

0.10 0.06 critical exponents are almost identical for both order parameters, in agreement with the property of the DQCP67. 0.03 The final results are β = 0.53(3) and ν = 1.55(6). In- 0.00 terestingly, we find the quantity 2ν(1 2β/ν) equals to 0.05 1 roughly, as predicted from the Luttinger− theory where both order parameters could be expressed by a sole Lut- tinger parameter67,68. We also check the cases where K = 1/226 and indeed find that the critical exponents 0.00 6 1.1 change with K. Together with the central charge c 1, ≈ 0.9 we could say that the VBS–zFM transition belongs to the

0.8 0.7 Gaussian universality class and the critical point shows some similarities to the LL phase. 0.6 0.9

0.4 TABLE I: Extracted critical point Jz,c and corresponding critical exponents β and ν for the continuous VBS–zFM phase transition. 0.0 1.2 1.4 1.6 1.8 Phase Jz,c β/ν 1/ν β ν 2(ν − 2β) VBS 1.4647(3) 0.344(2) 0.64(2) 0.53(2) 1.55(5) 0.98 zFM 1.4645(5) 0.350(3) 0.65(3) 0.54(2) 1.54(7) 0.92 FIG. 5: (a) The first two energy gaps ∆1 (open symbols) and ∆2 (filled symbols). The inset shows a linear extrapolation S of ∆2 to TDL. (b) Evolution of vNE L. Inset: Logarithmic Conclusions.– In this paper, we propose a bond re- S extrapolation of peaks of L at different length L’s. versal method to determine a first-order quantum phase transition (QPT) by a quantity called the difference of bond strength (DBS). A first-orderD QPT could be de- the correlation length and β is the critical exponent of tected by a jump in DBS and the discontinuity point is the order parameter such that M J J β near the ∼ | z − z,c| exactly the transition point. The method is rather effi- critical point Jz,c. cient and could be easily implemented in almost every numerical methods. We use it to study two unconven- 1.5 2.5 tional QPTs which are both beyond the scope of Landau- Ginzburg-Wilson theory. For the cross-coupled (J×) spin 2.0 ladder, we clarify that a continuous QPT indeed occurs at 1.0 weak interchain couplings, and the inflection point sep- 1.5 arating the continuous and first-order QPTs is J×,I 0.30(2). For a recently proposed spin-1/2 chain which≃ 1.0 0.5 owns two spontaneously symmetry breaking phases, we confirm that the transition is continuous because the 0.5 DBS is fairly smooth and the energy gap vanishes when 0.0 0.0 crossing the critical point. After a careful finite-size scal- -10 0 10 -10 0 10 ing analysis, we find that the transition belongs to the Gaussian universality class with the central charge c = 1. VBS Acknowledgements.– We thank S. Hu and S. Jiang for FIG. 6: The FSS of the order parameters (a) ML and zFM (b) ML . The critical point Jz,c and critical exponents β fruitful discussions. Q.L. was financially supported by and ν are shown in Tab. I. the Outstanding Innovative Talents Cultivation Funded Programs 2017 of Renmin University of China. J.Z. was In Fig. 6 we apply the finite-size scaling (FSS) method supported by the the National Natural Science Founda- to the (a) VBS and (b) zFM phases in the range of tion of China (Grant No. 11874188) and the Fundamen- J [1.20, 1.80]. The scaling results are pretty good tal Research Funds for the Central Universities. X.W. z ∈ when Jz is close to the critical point Jz,c. Some data, was supported by the National Program on Key Research however, deviate from the scaling function when Jz is far Project (Grant No. 2016YFA0300501) and by the Na- away from Jz,c. The best fitting values of critical point tional Natural Science Foundation of China (Grant No. Jz,c and critical exponents are presented in Tab. I. The 11574200). overall critical point Jz,c 1.4646(6), which is fairly in Note added.–Recently, we became aware of a work on consistent with previous work≃ by Huang et. al 69. The spin-1/2 chain with DQCP that supports our findings75. 6

∗ [email protected] 30 U. Schollwöck, Rev. Mod. Phys. 77, 259 (2005). † [email protected] 31 U. Schollwöck, J. Richter, D. J. J. Farnell, and R. F. ‡ [email protected] Bishop, Quantum Magnetism. (Springer Berlin Heidel- 1 S. Sachdev, Quantum Phase Transitions (Cambridge Uni- berg, 2004). versity Press, Cambridge, 2011). 32 C. N. Yang and C. P. Yang, Phys. Rev. 150, 321 (1966). 2 C. Lacroix, P. Mendels, and F. Mila, 33 M. Shiroishi and M. Takahashi, J. Phys. Soc. Jpn. 74, 47 Introduction to Frustrated Magnetism (Springer-Verlag (2015). Berlin Heidelberg, 2011). 34 J. D. Cloizeaux and M. Gaudin, J. Math. Phys. 7, 1384 3 B. Zeng, X. Chen, D.-L. Zhou, and X.-G. Wen, (1966). Quantum Information Meets Quantum Matter (Springer- 35 M. Bortz, J. Sato, and M. Shiroishi, J. Phys. A: Math. Verlag New York, 2019). Theor. 40, 4253 (2007). 4 J. M. Kosterlitz, and D. J. Thouless, J. Phys. C 6, 1181 36 M. Jimbo and T. Miwa, J. Phys. A 29, 2923 (1996). (1973). 37 G. Kato, M. Shiroishi, M. Takahashi, and K. Sakai, J. 5 D. C. Tsui, H. L. Stormer, and A. C. Gossard, Phys. Rev. Phys. A 36, L337 (2003). Lett. 48, 1559 (1982). 38 L. Banchi, F. Colomo, and P. Verrucchi, Phys. Rev. A 80, 6 F. D. M. Haldane, Phys. Rev. Lett. 50, 1153 (1983). 022341 (2009). 7 T. Senthil, A. Vishwanath, L. Balents, S. Sachdev, and M. 39 E. Ercolessi, S. Evangelisti, F. Franchini, and F. Ravanin, P. A. Fisher, Science 303, 1490 (2004). Phys. Rev. B 83, 012402 (2011). 8 T. Senthil, L. Balents, S. Sachdev, A. Vishwanath, and M. 40 V. Alba, M. Haque, and A. M Läuchli, J. Stat. Mech. 2012, P. A. Fisher, Phys. Rev. B 70, 144407 (2004). P08011 (2012). 9 A. W. Sandvik, Phys. Rev. Lett. 98, 227202 (2007). 41 J. Stasinska, B. Rogers, M. Paternostro, G. De Chiara, and 10 H. Shao, W. Guo, A. W. Sandvik, Science 352, 213 (2016). A. Sanpera, Phys. Rev. A 89, 032330 (2014). 11 F. Pollmann, A. M. Turner, E. Berg, and M. Oshikawa, 42 P. Pfeuty, Ann. Phys. (N.Y.) 57, 79 (1970). Phys. Rev. B 81, 064439 (2010). 43 A. Dutta, G. Aeppli, B. K. Chakrabarti, U. Divakaran, 12 Q. Faure, S. Takayoshi, S. Petit, V. Simonet, S. Raymond, T. F. Rosenbaum, and D. Sen, Quantum Phase Transitions L.-P. Regnault, M. Boehm, J. S. White, M. M˚ansson, C. in Transverse Field Spin Models. (Cambridge University Rüegg, P. Lejay, B. Canals, T. Lorenz, S. C. Furuya, T. Press, 2015). Giamarchi, and B. Grenier, Nat. Phys. 14, 716 (2018). 44 R. Coldea, D. A. Tennant, E. M. Wheeler, E. Wawrzynska, 13 X. Wang, Modern Phys. Lett. B 14, 327 (2000). D. Prabhakaran, M. Telling, K. Habicht, P. Smeibidl, and 14 A. Amaricci, J. C. Budich, M. Capone, B. Trauzettel, and K. Kiefer, Science 327, 177 (2010). G. Sangiovanni, Phys. Rev. Lett. 114, 185701 (2015). 45 A. B. Zamolodchikov, Int. J. Mod. Phys. A 4, 4235 (1989). 15 S. Barbarino, G. Sangiovanni, and J. C. Budich, 46 A. A. Ovchinnikov, D. V. Dmitriev, and V. Ya. Krivnov, Phys. Rev. B 99, 075158 (2019). and V. O. Cheranovskii, Phys. Rev. B 68, 214406 (2003). 16 B. Roy, P. Goswami, and J. D. Sau, Phys. Rev. B 94, 47 Y. Y. Atas and E. Bogomolny, J. Phys. A: Math. Theor. 041101(R) (2016). 50, 385102 (2017). 17 A. W. Sandvik, Phys. Rev. Lett. 104, 177201 (2010). 48 J. G. Bednorz and K. A. Müller, Z. Phys. B 64, 189 (1986). 18 K. Chen, Y. Huang, Y. Deng, A. B. Kuklov, N. V. 49 Y. Xian, Phys. Rev. B 52, 12485 (1995). Prokof’ev, and B. V. Svistunov, Phys. Rev. Lett. 110, 50 S. Wessel, B. Normand, F. Mila, and A. Honecker, SciPost 185701 (2013). Phys. 3, 005 (2017). 19 A. Iaizzi, K. Damle, and A. W. Sandvik, Phys. Rev. B 98, 51 E. Dagotto and T. M. Rice, Science 271, 618 (1996). 064405 (2018). 52 S. R. White, Phys. Rev. B 53, 52 (1996). 20 J. M. Fink, A. Dombi, A. Vukics, A. Wallraff, and P. 53 T. Vekua and A. Honecker, Phys. Rev. B 73, 214427 Domokos, Phys. Rev. X 7, 011012 (2017). (2006). 21 C. R. Laumann, R. Moessner, A. Scardicchio, and S. L. 54 A. Metavitsiadis and S. Eggert, Phys. Rev. B 95, 144415 Sondhi, Phys. Rev. Lett. 109, 030502 (2012). (2017). 22 M. Mueller, W. Janke, and D. A. Johnston, 55 K. Hijii, A. Kitazawa, and K. Nomura, Phys. Rev. B 72, Phys. Rev. Lett. 112, 200601 (2014). 014449 (2005). 23 M. Campostrini, J. Nespolo, A. Pelissetto, and E. Vicari, 56 Z. Weihong, V. Kotov, and J. Oitmaa, Phys. Rev. B 57, Phys. Rev. Lett. 113, 070402 (2014). 11439 (1998). 24 A. Yuste, C. Cartwright, G. De Chiara, and A. Sanpera, 57 E. H. Kim, G. Fáth, J. Sólyom, and D. J. Scalapino, New J. Phys. 20, 043006 (2018). Phys. Rev. B 62, 14965 (2000). 25 D. Rossini and E. Vicari, Phys. Rev. E 98, 062137 (2018). 58 E. H. Kim, O.¨ Legeza, and J. Sólyom, Phys. Rev. B 77, 26 For details see the Supplemental Material at [url], which 205121 (2008). includes the DBS of continuous QPTs, QPTs of the cross- 59 O. A. Starykh and L. Balents, Phys. Rev. Lett. 93, 127202 coupled spin ladder, and the critical exponents of the spin- (2004). 1/2 chain with DQCP. 60 G.-H. Liu, H.-L. Wang, and G.-S. Tian, Phys. Rev. B 77, 27 S. R. White, Phys. Rev. Lett. 69, 2863 (1992). 214418 (2008). 28 S. R. White, Phys. Rev. B 48, 10345 (1993). 61 Y.-C. Li and H.-Q. Lin, New J. Phys. 14, 063019 (2012). 29 I. Peschel, X. Q. Wang, M. Kaulke, and K. Hallberg, 62 H.-H. Hung, C.-D. Gong, Y.-C. Chen, and M.-F. Yang, Density-matrix renormalization. (Springer Berlin Heidel- Phys. Rev. B 73, 224433 (2006). berg, 1999). 63 E. H. Kim, O.¨ Legeza, and J. Sólyom, Phys. Rev. B 77, 7

205121 (2008). 64 T. Hikihara and O. A. Starykh, Phys. Rev. B 81, 064432 (2010). 65 G. Barcza, O.¨ Legeza, R. M. Noack, and J. S´olyom, Phys. Rev. B 86, 075133 (2012). 66 X.-H. Chen, S. Y. Cho, H.-Q. Zhou, and M. T. Batchelor, J. Korean Phys. Soc. 68, 1114 (2016). 67 S. Jiang and O. Motrunich, Phys. Rev. B 99, 075103 (2019). 68 B. Roberts, S. Jiang, and O. I. Motrunich, Phys. Rev. B 99, 165143 (2019). 69 R.-Z. Huang, D.-C. Lu, Y.-Z. You, Z. Y. Meng, and T. Xiang, arXiv:1904.00021 (2019). 70 C. Mudry, A. Furusaki, T. Morimoto, and T. Hikihara, Phys. Rev. B 99, 205153 (2019). 71 C. K. Majumdar and D. K. Ghosh, J. Math. Phys. 10, 1388 (1969); ibid, 10, 1399 (1969). 72 P. Di Francesco, P. Mathieu, and D. Senechal, Conformal ﬁeld theory (Springer, New York, 1997). 73 P. Calabrese, and J. Cardy, J. Stat. Mech.: Theor. Exp. 2004, P06002 (2004). 74 M. N. Barber, Phase Transitions and Critical Phenomena Vol. 8 (eds C. Domb and J. L. Leibovitz) (Academic, Lon- don, 1983). 75 G. Sun, B.-B. Wei, and S.-P. Kou, arXiv:1906.03850 (2019). 1

Supplemental Material for “Intrinsic Jump Character of the First-Order Quantum Phase Transitions”

Qiang Luo1, Jize Zhao2, and Xiaoqun Wang3,4,5

1Department of Physics, Renmin University of China, Beijing 100872, China 2School of Physical Science and Technology & Key Laboratory for Magnetism and Magnetic Materials of the MoE, Lanzhou University, Lanzhou 730000, China 3Key Laboratory of Artiﬁcial Structures and Quantum Control (Ministry of Education), School of Physics and Astronomy, Tsung-Dao Lee Institute, Shanghai Jiao Tong University, Shanghai 200240, China 4Collaborative Innovation Center for Advanced Microstructures, Nanjing 210093, China 5Beijing Computational Science Research Center, Beijing 100084, China

I. DBS OF CONTINUOUS QPTS 0.25 -0.3 A. From the second-order Ising transition to the 0.20 infinite-order KT transition -0.4 0.15 In this section we show the difference of bond strength (DBS) D of continuous QPTs of the KT (infinite order) -0.5 0.10 and Ising (second order) universality classes, respectively. For the spin-1/2 XXZ Heisenberg chain, the transition 0.05 -0.6 from the Luttinger liquid phase to the AFM phase occurs at ∆ = 1 and belongs to the KT universality classes. 0.00 In the critical region, i.e, 1 < ∆ 1, the correlation 0 1 2 0 1 2 functions read as1 − ≤ 1 cos πυ sm SxSx = + (sm-1) FIG. -1: (a) The ground-state energy eg of XXZ chain for h i i+1i −4π sin πυ I1 4π2 I2 L = 64 (red square), 128 (blue circle), and TDL (black line). (b) The same setup as (a) for DBS DL. and 1 cot πυ 1 -1.00 1.0 SzSz = + , (sm-2) h i i+1i 4 2π I1 − 2π2 I2 -1.25 where ∆ = cos πυ and the integrals 0.5 ∞ dx sinh(1 υ)x -1.50 1 1(υ)= − 0.0 I ≡ I sinh x cosh υx Z−∞ -1.75 and -0.5 -2.00 ∞ dx x cosh x 2 2(υ)= 2 . I ≡ I −∞ sinh x (cosh υx) -2.25 -1.0 Z 0 1 2 0 1 2 In the massive region, i.e, ∆ = cosh φ> 1, the correlation functions are2

∞+i/2 FIG. sm-2: (a) The ground-state energy eg of TFIM for L = x x 1 dx x sinh 2φ sin 2φx 64 (red square), 128 (blue circle), and TDL (black line). (b) Si Si+1 = 2 − h i 8 sinh πx sin φx sinh φ The same setup as (a) for DBS DL. Z−∞+i/2 (sm-3) and For the TFIM, there is an Ising transition at hz = 1 1 ∞+i/2 dx sin 2φx coth φ 2x 1 that separates the FM phase from the paramagnetic SzSz = + − . h i i+1i 4 4 sinh πx sin2 φx phase. The energy per site is3 Z−∞+i/2 (sm-4) z z 2 2√hz The DBS is defined as L = S S e = (1 + h )E (sm-5) D h L/2 L/2+1i − g −π z 1+ h 2 Sx Sx . and the final result is presented in z h L/2 L/2+1i Fig. sm-1. It could be found that the DBS is smooth where E( ) is the complete elliptic integral of the second when crossing the critical point ∆ = 1. kind. The· transverse magnetization M = σz could z h i i 2 also be calculated as4 II. QPTS OF THE CROSS-COUPLED SPIN 2 LADDER (1−hz) 2 Mz πhz Π(hz; hz) K(hz) , hz < 1 = 2 − . (sm-6) (hz−1) 1 1 2 2 Π 2 ; , h > 1 A. Continuous QPT at J× = 0.2 ( πhz hz hz z x x The correlation function σi σi+1 = (eg + hzMz). The The fact that a direct continuous QPT occurs at J h x x i −z × DBS is defined as L = σi σi+1 σi , and the final re- = 0.2 has been checked by several different methods. D hsm i−h i sult is presented in Fig. -2. It could be found that the We here revisit the problem by studying the energy gap DBS is smooth when crossing the critical point hz = 1. and central charge. Before carrying large-scale numeri- Since the Ising transition is the continuous QPT of the cal calculations, we firstly present some details about our second order and the KT transition is the continuous DMRG simulations to show the convergence of our re- QPT of the infinite order, we thus could arrive at the sults. In Fig. sm-4 we show the behavior of ground-state conclusion that the DBS is always smooth (or continu- energy Eg versus states kept m. It could be found that ous) for the continuous QPT. As a result, the jump of as long as 2000 states are kept in the simulations we shall DBS is a characteritic feature of the first-order QPT. (m) (∞) reduce the absolute error of energy ǫE = Eg Eg to (m) (∞)− seven place of decimals where Eg and Eg represent B. Commensurate-incommensurate transition the energy at current states kept m and infinite states kept m . The truncated error of information loss We now consider the spin-1/2 XXZ chain under the here is of→ the ∞ order 10−12, which is fairly small. There- longitudinal field. The Hamiltonian reads as5 fore, we keep typically 2000 states in our calculations and 4-8 sweeps are performed to ensure our results are well = J (SxSx + SySy + ∆SzSz ) h Sz H i i+1 i i+1 i i+1 − z i converged. i i X X 4 -1 where J = 1 and hz ( 0) is the longitudinal field. There are two gapped phases:≥ A ferromagnetic one at suffi- ciently strong fields and an antiferromagnetic phase for 3 -3 ∆ > 1 at small fields in the full phase diagram. Also, a massless Luttinger phase is sandwiched between the two6. The longitudinal correlation function of the Lut- 2 -5 tinger phase is incommensurate with sinusoidally mod- 7 ulated behavior . The transition between the ferromag- 1 -7 netic commensurate phase and the massless incommensurate phase, which occurs on the line hu,c/J = 1 + ∆, is an example of the Dzhaparidze-Nersesyan-Pokrovsky- 0 -9 Talapov universality class8,9. We now focus on the line 0 1k 2k 3k 0 1k 2k 3k x x of ∆ = 2, and we define the DBS as L = (Si Si+1 + SySy +2SzSz ) Sz . The finalD resulth is pre- i i+1 i i+1 − i ii=L/2 sented in Fig. sm-3. It could be found that the DBS is FIG. sm-4: Evolution of (a) the absolute error of energy ǫE continuous when crossing the critical point hz,c = 3. and (b) its logarithmic form versus states kept m. Here, total length of the ladder is L = 128 and J× = 0.2. Two different -0.8 points at J⊥ = 0.30 (red square) and 0.40 (blue circle) are selected as the representative points in the Haldane phase 0.50 -0.9 and the RS phase, respectively.

-1.0 We now turn to calculate the energy gap ∆ at the crit- 0.25 ical point and the central charge c, if any. In the open -1.1 boundary condition, the Haldane phase has a four-fold degenerate ground state due to the edge modes. The Hal- -1.2 0.00 dane gap ∆L is thus defined as the difference of ground- z z state energy in the Stot = 2 and Stot = 0 subspaces, i.e., -1.3 ∆ = E Sz =2 E Sz =0 . The DMRG result of L g tot − g tot 2.5 3.0 3.5 2.5 3.0 3.5 the Haldane gap ∆L is shown in Fig. sm-5 and it could be seen that local minima exist near the critical point in the finite systems. Such minimal gaps obey a linear scaling sm FIG. -3: (a) The ground-state energy eg for L = 64 (cyan formula and the gap in the TDL is 0.0001(2), which could rhombus), L = 128 (magenta square), and L = 192 (blue be regarded as zero within the numerical precision. The D triangle). (b) The same setup as (a) for DBS L. vanishing gap at the critical point is a typical signal for a continuous QPT, and the linear fitting suggests that the 3

critical point could be described by the conformal field B. First-Order QPT at J× = 0.4 theory. Whereas the DBS of the cross-coupled spin ladder 0.3 0.15 seems to be smooth when J× = 0.4 (see Fig. 3 in the 0.10 main text), we want to convince the readers that there is a jump actually. For this purpose we calculate the DBS 0.05 0.2 of a longer ladder whose length L is up to 512. As shown 0.00 sm 0.00 0.02 in Fig -7, we find that the energy curves are more and more screwy and a kink is expected for an infinite system. 0.1 Likewise, for the DBS, it is very smooth for small sizes and a jump could be spotted when L = 512. Therefore, we think that the transition at J× = 0.4 is still of first order and there is a jump of DBS correspondingly. 0.0 0.30 0.35 0.40 0.45 -0.894 0.20

FIG. sm-5: The Haldane gap ∆L at ﬁxed J⊥ = 0.2 for diﬀer- ent chain length L. The inset shows a linear extrapolation of -0.895 ∆L to TDL. 0.15

2.6 0.10 -0.897

2.2 -0.898 0.05 0.70 0.71 0.72 0.70 0.71 0.72 2.6 1.8

2.3 FIG. sm-7: (a) The ground-state energy eg of the cross- coupled spin ladder for L = 128 (magenta square), L = 1.4 192 (blue triangle), L = 256 (red circle), L = 320 (cyan rhom- 2.0 0.7 0.9 bus), and L = 512 (auburn star) at ﬁxed J× = 0.4. (b) The same setup as (a) for DBS DL. 1.0 0.2 0.3 0.4 0.5

sm S FIG. -6: The von Neumann entropy L at fixed J⊥ = 0.2 0.15 for different chain length L. The inset shows a proper fitting 0.3 of the maxima of SL to extract the central charge c and the 0.10 best estimated value is c = 1.98(4). 0.05 0.2 0.00 Meanwhile, we also calculate the von Neumann en- 0.00 0.02 tropy L and the result is presented in Fig. sm-6. A 0.1 bump couldS be spotted near the critical point and is consistent with a continuous QPT. It is well established that for the critical system under the OBC the vNE obeys the 0.0 following formula, 0.60 0.65 0.70 0.75

c 2L πl ′ sm L(l)= ln sin + c (sm-7) FIG. -8: The Haldane gap ∆L at fixed J× = 0.4 for differ- S 6 π L ent chain length L. The inset shows a quadratic extrapolation of ∆L to TDL. where c is the central charge. The result in Fig. sm-6 is a special case where l = L/2. It could be seen from the inset that the central charge c 2. It’s worth mention- Likewise, we calculate the Haldane gap as before and ing that the central charge at the≃ decoupling limit where find that it is finite of 0.0056(8) in the TDL. This further J⊥ = 0 is also equal to 2. confirms that the transition at J× = 0.4 is of first order. 4

C. Transition points and the Haldane gaps ical point Kc = 0.5497(1) and the associated critical exponents β = 0.21(1) and ν = 0.92(3). Those results are 10 The transition points are obtained by the joint analysis in fairly consistent with other group . Together with of the maxima of vNE and minima of Haldane gap the result shown in the main text, we can conclude that SL ∆L for the continuous QPT, and by the jump of DBS the critical exponents vary along the locus of the phase for the ﬁrst-order one. The ﬁnal results are presented in boundary. Tab. sm-1.

sm TABLE -1: Transition point J⊥,t and triplet gap ∆m of 0.5 Haldane phase thereof for several selected J×.

J× J⊥,t ∆m J× J⊥,t ∆m 0.4 0.20 0.3826(3) 0.0001(2) 0.45 0.7918(6) 0.0158(2) 0.30 0.5576(5) 0.0004(5) 0.50 0.8625(5) 0.0355(3) 0.3 0.40 0.7168(3) 0.0037(9) 0.60 0.9990(5) 0.1265(5) 0.2

0.1 III. CRITICAL EXPONENTS OF THE SPIN-1/2 CHAIN 0.0 0.54 0.55 0.56 -5 0 5 In this section we pick up another set of parameters shown in Eq. (3) of the main text. Here we fix FIG. sm-9: FSS analysis of the VBS order parameter ML for ˜ J = (Jz + Jx)/2 = 1 as the energy unit and introduce δ = 0.5 as a function of K/(Jx +Jz). The best fitting suggests an anisotropic parameter δ = (Jz Jx)/(Jz + Jx). Let Kc = 0.5497(1) and critical exponents β = 0.21(1) and ν = δ = 0.5 and the finite-size scaling− (FSS) analysis of the 0.92(3). corresponding VBS order parameter ML are shown in Fig. sm-9. After a careful analysis, we find that the crit-

∗ [email protected] Bishop, Quantum Magnetism. (Springer Berlin Heidel- † [email protected] berg, 2004). ‡ [email protected] 7 B. Grenier, V. Simonet, B. Canals, and P. Lejay, M. Klan- 1 G. Kato, M. Shiroishi, M. Takahashi, and K. Sakai, J. jsek, M. Horvatic, and C. Berthier, Phys. Rev. B 92, Phys. A 36, L337 (2003). 134416 (2015). 2 M. Takahashi, G. Kato, and M. Shiroishi, J. Phys. Soc. 8 G. I. Dzhaparidze, A. A. Nersesyan, JETP Lett. 27, 334 Jpn. 73, 245 (2004). (1978). 3 P. Pfeuty, Ann. Phys. (N.Y.) 57, 79 (1970). 9 V. L. Pokrovsky, A. L. Talapov, Phys. Rev. Lett. 42, 65 4 T. Maciazek and J. Wojtkiewicz, Physica A 441, 131 (1979). (2016). 10 B. Roberts, S. Jiang, and O. I. Motrunich, Phys. Rev. B 5 C. N. Yang and C. P. Yang, Phys. Rev. 150, 321 (1966). 99, 165143 (2019). 6 U. Schollw¨ock, J. Richter, D. J. J. Farnell, and R. F.