Quantifying the Phase Diagram and Hamiltonian of S = 1 / 2 S=1/2 Kagome Antiferromagnets: Bridging Theory and Experiment Shengtao Jiang (gbsn蒋晟韬), Arthur C. Campello, Wei He, Jiajia Wen, Daniel M. Pajerowski, Young S. Lee, Hong-Chen Jiang Abstract Spin- 1 / 2 1/2 kagome antiferromagnets are leading candidates for realizing quantum spin liquid (QSL) ground states. While QSL ground states are predicted for the pure Heisenberg model, understanding the robustness of the QSL to additional interactions that may be present in real materials is a forefront question in the field. Here we employ large-scale density-matrix renormalization group simulations to investigate the effects of next-nearest neighbor exchange couplings J 2 J_{2} and Dzyaloshinskii-Moriya interactions D D , which are relevant to understanding the prototypical kagome materials herbertsmithite and Zn-barlowite. By utilizing clusters as large as XC12 and extrapolating the results to the thermodynamic limit, we precisely delineate the scope of the QSL phase, which remains robust across an expanded parameter range of J 2 J_{2} and D D . Direct comparison of the simulated static and dynamic spin structure factors with inelastic neutron scattering reveals the parameter space of the Hamiltonians for herbertsmithite and Zn-barlowite, and, importantly, provides compelling evidence that both materials exist within the QSL phase. These results establish a powerful convergence of theory and experiment in this most elusive state of matter. Introduction. Quantum spin liquids (QSLs) are exotic phases of matter that avoid spontaneous symmetry breaking even at zero temperature and support fractionalized excitations Anderson 1973 ; Kivelson et al. 1987 ; Rokhsar and Kivelson 1988 ; Laughlin 1988 ; Kivelson 1989 ; Moessner and Sondhi 2001 ; Balents 2010 ; Savary and Balents 2016 ; Broholm et al. 2020 . A prominent system known to host a QSL is the kagome lattice antiferromagnet, characterized by strong quantum fluctuations induced by pronounced geometric frustration. Indeed, a consensus has been reached that the kagome antiferromagnetic (AF) Heisenberg model with nearest-neighbor (NN) exchange coupling J_{1} has a QSL ground state, although its precise nature, i.e., gapped (such as a Z_{2} QSL) or gapless (such as a U(1) Dirac QSL), remains under debate Jiang et al. 2008 ; Yan et al. 2011 ; Jiang et al. 2012 ; Depenbrock et al. 2012 ; Mei et al. 2017 ; Ran et al. 2007 ; Liao et al. 2017 ; He et al. 2017 ; Gong et al. 2015 ; Haghshenas et al. 2019 ; Zhu et al. 2019 ; Läuchli et al. 2011 ; Läuchli et al. 2019 ; Sun et al. 2024 ; Iqbal et al. 2013 ; Iqbal et al. 2014 ; Sindzingre and Lhuillier 2009 . On the experimental side, significant progress has been made in the synthesis and investigation of QSL candidate materials. A leading example is herbertsmithite ZnCu 3 (OH) 6 Cl 2 Helton et al. 2007 ; Mendels et al. 2007 ; Han et al. 2012a ; Norman 2016 , in which the spin-1/2 moments on Cu 2+ are arranged on a structurally perfect kagome lattice, with nonmagnetic Zn 2+ ions separating those kagome planes. Experiments indicate a dominant NN AF coupling J_{1}\approx 17 meV in herbertsmithite Helton et al. 2007 ; Han et al. 2012a , which establishes it as an ideal platform for realizing and exploring the QSL state. This is indeed evidenced by the experimental observations demonstrating the absence of magnetic order down to temperatures as low as 50 mK Helton et al. 2007 ; Mendels et al. 2007 ; de Vries et al. 2009 . Zn-barlowite ( \text{ZnCu}{\vphantom{\text{X}}}_{\smash[t]{\text{3}}}\text{(}\text{OH}\text{)}\text{}{\vphantom{\text{X}}}_{\smash[t]{\text{6}}}\text{FBr} ) is another well-established kagome QSL candidate material Feng et al. 2017 ; Smaha et al. 2018 ; Smaha et al. 2020 ; Tustain et al. 2020 ; Fu et al. 2021 with a dominant J_{1}\approx 14 meV Smaha et al. 2020 . For both materials, fractionalized spin excitations have been observed in inelastic neutron scattering (INS) measurements on single crystal samples consistent with a QSL state Han et al. 2012a ; Han et al. 2016 ; Breidenbach et al. 2025 ; Campello 2025 ; Khuntia et al. 2020 . Furthermore, experiments suggest a finite spin gap for the intrinsic kagome moments in both materials Han et al. 2016 ; Fu et al. 2015 ; Wang et al. 2021 ; Breidenbach et al. 2025 , which may indicate a Z_{2} QSL. Despite significant recent theoretical and experimental progress, further steps are necessary to bridge the gap between the QSL in the J_{1} -only kagome AF model and real materials where additional interactions may also be present. Indeed, it has been proposed that a minimal spin model for the two kagome materials include a small but finite next-nearest-neighbor (NNN) Heisenberg interaction J_{2} Han et al. 2012a ; Jeschke et al. 2013 . Meanwhile, the bulk magnetic properties, anisotropies in thermodynamic quantities, and electron spin resonance measurements suggest the presence of a weak Dzyaloshinskii-Moriya (DM) interaction Zorko et al. 2008 ; Han et al. 2012b ; Rigol and Singh 2007 ; Messio et al. 2010 , which arises from spin-orbit coupling Lee and Nagaosa 2013 . The effects of the J_{2} or DM interaction have been explored in various contexts using semi-classical or numerical approaches Harris et al. 1992 ; Chubukov 1992 ; Elhajal et al. 2002 ; Huh et al. 2010 ; Cépas et al. 2008 ; Chernyshev and Zhitomirsky 2015 ; Gong et al. 2015 ; Kolley et al. 2015 . In particular, it is crucial to investigate how these additional interactions influence the stability of the QSL state and establish the microscopic minimal model for herbertsmithite and Zn-barlowite on a quantitative level. In this Letter, we address these questions and bridge the gap between our understanding of QSL’s in models and real materials using large-scale density-matrix renormalization group (DMRG) White 1992 ; White 1993 simulations with input from neutron scattering measurements. We establish the ground-state phase diagram of the kagome AFM model in the J_{2} and D plane, demonstrating that the QSL phase remains robust in the presence of both interactions. Specifically, the QSL phase is stable within the ranges -0.07\lesssim J_{2}\lesssim 0.18 and 0\lesssim D\lesssim 0.06 as shown in Fig. 1 (c). By comparing both the static and dynamic spin structural factors with neutron scattering results, we find that the model parameters for both materials are constrained to similar regions, and both fall entirely within the QSL phase. Model. – The kagome lattice Heisenberg J_{1} - J_{2} model with a DM interaction, depicted in Fig. 1 (a), is defined as: H=J_{1}\sum_{\langle ij\rangle}\mathbf{S}_{i}\cdot\mathbf{S}_{j}+J_{2}\sum_{\langle\langle ij\rangle\rangle}\mathbf{S}_{i}\cdot\mathbf{S}_{j}+\sum_{\langle ij\rangle}\mathbf{D}_{ij}\cdot(\mathbf{S}_{i}\times\mathbf{S}_{j}). (1) Here \mathbf{S}_{i} is the S=1/2 spin operator on site i , and the first two terms denote spin exchange couplings between NN and NNN sites, respectively. The third term is the DM interaction originating from spin-orbit coupling, which can be present when lattice inversion symmetry is broken Dzyaloshinsky 1958 ; Moriya 1960 ; Lee and Nagaosa 2013 . The DM vector \mathbf{D}_{ij} depends on the convention of the bond orientation as \mathbf{D}_{ij}=-\mathbf{D}_{ji} Elhajal et al. 2002 ; Cépas et al. 2008 ; Zorko et al. 2008 . For a given convention Lee and Nagaosa 2013 , the DM vectors are shown in Fig. 1 (a). Since electron-spin resonance measurements suggest the existence of a leading out-of-plane component of \mathbf{D}_{ij} Zorko et al. 2008 , we hence choose \mathbf{D}_{ij}=D\hat{z} in the present study. This reduces the computational costs and facilitates reliable DMRG simulations on large systems by leveraging the U(1) spin rotational symmetry. We adopt the convention where D>0 corresponds to all bonds i\rightarrow j being oriented clockwise, as illustrated in Fig. 1 (a) Lee and Nagaosa 2013 . We set J_{1}=1 as an energy unit. In the absence of J_{2} and DM interactions, numerous numerical studies have consistently identified a QSL ground state Jiang et al. 2008 ; Yan et al. 2011 ; Jiang et al. 2012 ; Depenbrock et al. 2012 ; Mei et al. 2017 ; Ran et al. 2007 ; Liao et al. 2017 ; He et al. 2017 ; Sun et al. 2024 ; Iqbal et al. 2013 ; Iqbal et al. 2014 ; Sindzingre and Lhuillier 2009 ; Gong et al. 2015 , although its precise nature, i.e., gapped or gapless in the thermodynamic limit, remains currently under debate. We do not aim to further address the gapped versus gapless nature of the QSL, as this question is beyond the scope of the current study. We are interested in examining the robustness of the QSL state, where previous DMRG studies suggest that the QSL remains the ground state in an extended parameter region between J_{2}\leq 0.15\sim 0.20 and J_{2}\geq-0.1\sim-0.05 in the absence of DM interaction Gong et al. 2015 ; Kolley et al. 2015 . However, the ground state phase diagram in the presence of both J_{2} and DM interactions remains largely unexplored. Moreover, the effects of these parameters on the spin excitations of herbertsmithite and Zn-barlowite have yet to be quantitatively explored. This study aims to address these questions by simulating the spin Hamiltonian in Eq.( 1 ) and comparing the results with results from inelastic neutron scattering. Figure 1: (a) Heisenberg J_{1} - J_{2} - D model in Eq.( 1 ) on a kagome cylinder. Periodic and open boundary conditions are imposed, respectively, along the directions specified by the lattice basis vectors e_{2} and e_{1} . The small triangle in the shaded region denotes a unit cell. J_{1} and J_{2} are NN and NNN spin exchange couplings, and D is the DM interaction. (b) The first and extended Brillouin zones, the reciprocal lattice vectors b_{1} and b_{2} , and the high-symmetry points: \Gamma , \Gamma^{\prime} , \rm K^{\prime} , P (the midpoint of \Gamma^{\prime} , \rm K^{\prime} ), and G=5b_{2}/6 . (c) Ground state phase diagram of the system as a function of J_{2} and D . The solid symbols with error bars label the phase boundaries determined by DMRG calculations. Phase diagram. – Based on large-scale DMRG simulations, we establish the ground state phase diagram of the model Hamiltonian in Eq.( 1 ) as shown in Fig. 1 (c). The phases are determined by calculating the equal-time spin-spin correlation function and the corresponding spin structure factors. Our results reveal an extended QSL region that is characterized by diffuse structure factors without sharp peak-like features. Specifically, the QSL phase is stable within the range -0.07(1)\lesssim J_{2}\lesssim 0.18(1) in the absence of DM interaction, and D\lesssim 0.063(4) when J_{2} is absent. These results align well with previous studies along the J_{2} and D axis, respectively Kolley et al. 2015 ; Cépas et al. 2008 . Outside the spin liquid region, we find that the system develops long-range magnetic order of the \sqrt{3}\times\sqrt{3} type when J_{2} is more ferromagnetic and D is small. Conversely, for stronger AF J_{2} and/or larger D , the system transitions into the q=0 ordered state. The characteristic spin structure factors for both the magnetically ordered phases and QSL phase are presented in Fig. 2 . DMRG setup. – The lattice geometry used in our DMRG simulations is depicted in Fig. 1 (a), where \mathbf{e}_{1}=(2,0) and \mathbf{e}_{2}=(1,\sqrt{3}) denote the two basis vectors. We consider kagome cylinders with open and periodic boundary conditions along the \mathbf{e}_{1} and \mathbf{e}_{2} directions. Following the convention in Ref. Yan et al. 2011 ; Kolley et al. 2015 , we refer to a cylinder with L_{x} and L_{y} unit cells ( 2L_{y} and 2L_{x} sites) in the \mathbf{e}_{2} and \mathbf{e}_{1} directions as XC 2L_{y} - L_{x} . For a better connection to the two-dimensional (2D) limit White and Chernyshev 2007 , we primarily focus on “square-like” cylinders with aspect ratio 10 . These yield a typical truncation error \epsilon\lesssim 10^{-5} with excellent convergence for our results when extrapolated to the limit \epsilon=0 , i.e., m=\infty . Phase determination using static structure factor. – We begin with measurements of the equal-time spin-spin correlation \langle\mathbf{S}_{i}\cdot\mathbf{S}_{j}\rangle between sites i and j , and the corresponding static spin structure factor S(\mathbf{q}) , defined as \displaystyle S(\mathbf{q})=\frac{1}{N}\sum_{i,j=1}^{N}\langle\mathbf{S}_{i}\cdot\mathbf{S}_{j}\rangle e^{i\mathbf{q}\cdot(\mathbf{r}_{i}-\mathbf{r}_{j})}. (2) Figure 2 presents examples of S(\mathbf{q}) for the \sqrt{3}\times\sqrt{3} , QSL and \mathbf{q} =0 states, corresponding to characteristic parameter sets ( J_{2}=-0.1 , D=0 ), ( J_{2}=0 , D=0 ), and ( J_{2}=0.1 , D=0.1 ) on the XC12-9 cylinder. The structure factor S(\mathbf{q}) has a clear peak at \rm K^{\prime} momenta in the extended Brillouin zone (BZ) for ( J_{2}=-0.1 , D=0 ), corresponding to the long-ranged \sqrt{3}\times\sqrt{3} magnetic order. On the other hand, for ( J_{2}=0.1 , D=0.1 ), the structure factor is sharply peaked at \Gamma^{\prime} momenta (the center of the second Brillouin zone). This aligns well with the q=0 magnetic ordered state, which preserves the translational symmetry of the system. In stark contrast, S(\mathbf{q}) becomes diffuse without any sharp peak for ( J_{2}=0 , D=0 ), demonstrating the absence of any long-range magnetic order, consistent with a QSL state. Figure 2: Static spin structure factors S(\mathbf{q}) obtained from ground state DMRG simulations on XC12-9 cylinders for three characteristic sets of parameters: (a) J_{2}=-0.1 and D=0 , the system exhibits the \sqrt{3}\times\sqrt{3} magnetic order with sharp peaks in S(\mathbf{q}) at \rm K^{\prime} ; (b) J_{2}=0 and D=0 , the system has a QSL ground state with diffuse S(\mathbf{q}) ; (c) J_{2}=0.1 and D=0.1 , the system has q=0 order with sharp peaks in S(\mathbf{q}) at \rm\Gamma^{\prime} ; The first and extended Brillouin zones are indicated by the dashed hexagons. The color scale has an upper cutoff of 2. Figure 3: Squared \mathbf{q} =0 magnetic order parameter m^{2}(\mathbf{Q}=\Gamma^{\prime}) as a function of (a) J_{2} at D=0 and (b) D at J_{2}=0 . The shaded regions label the phase transitions between the QSL and \mathbf{q}=0 magnetically ordered phase. (c) Examples of finite-size extrapolations of m^{2}(\mathbf{Q}=\Gamma^{\prime}) for different D at J_{2}=0 using second-order polynomials in 1/\sqrt{N} . We use two different ways to quantify the magnetic order and determine the phase boundaries. We first define the squared order parameter as m^{2}(\mathbf{Q},N)=S(\mathbf{Q})/N , where \mathbf{Q} represents the peak positions associated with different magnetic orders, i.e., \mathbf{Q}={\rm K^{\prime}} for \sqrt{3}\times\sqrt{3} magnetic order and \mathbf{Q}={\Gamma^{\prime}} for q=0 magnetic order. Figure 3 (a) and (b) show m^{2}({\mathbf{Q}=\Gamma^{\prime}}) for the q=0 order as a function of J_{2} at D=0 and as a function of D at J_{2}=0 for different cylinders. To quantitatively analyze the order, we perform an extrapolation of m^{2}(\mathbf{Q},N) to the 2D limit as illustrated in Figure 3 (c), using results from various kagome cylinders ranging from XC6 to XC12. The extrapolated m^{2}(\mathbf{Q}) denotes the squared order parameter in the thermodynamic limit N=\infty , which remains finite in the magnetic ordered states but vanishes in the QSL state. Examples of finite-size extrapolations are shown in Fig. 3 (c) for various D at J_{2}=0 . It is evident that m^{2}(\mathbf{Q}=\Gamma^{\prime}) remains finite for D\gtrsim 0.07 , indicating the presence of the q=0 magnetic order. In contrast, for D\leq 0.06 , m^{2}(\mathbf{Q}=\Gamma^{\prime}) extrapolates to zero or even a negative value, indicating the absence of such an order. Consequently, the phase boundary between the q=0 and QSL phases is established at D\approx 0.06 along the J_{2} =0 line. Alternatively, the phase boundary can be identified by examining m^{2}(\mathbf{Q},N) and its derivative as functions of J_{2} or D . For a given system size, m^{2}(\mathbf{Q},N) remains relatively small in the QSL phase but becomes significantly larger in the magnetically ordered phases. Consequently, a sharp increase (or even a discontinuity in small systems) accompanied by a pronounced peak in its derivative is expected at the phase boundary as indicated by the shaded regions in Fig. 3 (a) and (b). Using this approach, we determine the phase boundaries between the QSL and the q=0 ordered phase to be around D=0.065(5) in the absence of J_{2} , and J_{2}=0.18(2) for D=0 . Similarly, the boundary between the QSL and the \sqrt{3}\times\sqrt{3} phase is found at J_{2}=-0.07(1) for J_{2}=0 . These phase boundaries determined in both ways are self-consistent and align well with previous studies Gong et al. 2015 ; Kolley et al. 2015 . Connection to kagome materials.– Having established the ground state phase diagram of the spin-1/2 Heisenberg J_{1} - J_{2} - D model on the kagome lattice, we then try to refine the parameters of the microscopic Hamiltonian for Zn-barlowite and herbertsmithite. This is achieved by comparing the static spin structure factor S(\mathbf{q}) from DMRG simulation with the neutron scattering experiments. Experimentally, the value of S(\mathbf{q}) is approximated by integrating S_{mag}(\mathbf{q},\omega) in the frequency region \omega=2.5 meV and 6 meV, which arise from scattering from the intrinsic kagome moments Han et al. 2016 ; Breidenbach et al. 2025 . To quantify the goodness of fit, we compare the quantity: \displaystyle\delta R(J_{2},D)=\sqrt{\sum_{i=1}^{3}(R_{i}(J_{2},D)-R_{i}^{E})^{2}}, (3) which represents the deviation between the DMRG simulations on the model Hamiltonian in Eq. ( 1 ) and the neutron data. The ratios of the structure factor S(\mathbf{q}) integrated around several high-symmetry momenta are selected. These include R_{1}=S({\rm\Gamma^{\prime}})/S({\rm K^{\prime}}) , R_{2}=S({\rm\Gamma^{\prime}})/S(\rm G) and R_{3}=S({\rm K^{\prime}})/S(\rm P) , as defined in Fig. 1 (b). Here, R_{i} and R_{i}^{E} are the ratios from DMRG simulations and neutron experiments, respectively. Figure 4: Contour plot of the deviation \delta R(J_{2},D) (defined in Eq. 3 ) between DMRG simulations as a function of J_{2} and D on XC12 cylinders and neutron scattering data on (a) Zn-barlowite Breidenbach et al. 2025 and (b) herbertsmithite Han et al. 2012a ; Campello 2025 . The white dashed lines denote phase boundaries in Fig. 1 . The best-fit regions of both materials fall within the QSL phase. The contour is plotted with interpolation and an upper cutoff of 1.5 in the color scale. Figure 4 shows the \delta R(J_{2},D) between DMRG simulation on XC12 kagome cylinders and the neutron measurement on Zn-barlowite ( R_{1} =1.126, R_{2} =1.430, R_{3} =1.130) Breidenbach et al. 2025 as well as herbertsmithite ( R_{1} =1.41, R_{2} =1.61, R_{3} =0.96) Han et al. 2012a ; Campello 2025 . There is a boundary at \delta R\approx 1.5 beyond which \delta R increases rapidly, enabling us to constrain the exchange parameters within this region, as shown in Fig. 6 . Within the boundaries, \delta R is significantly smaller and flatter. Moreover, experimental uncertainties, such as the energy integration window, can influence the details of this region. For instance, one would ideally integrate over all frequencies in the INS data. However, here we chose the frequency window of \omega=[2.5,6] meV, where non-intrinsic scattering from impurities and non-magnetic phonon scattering is minimized, as well as allowing for sufficient kinematic coverage in reciprocal space for integration Han et al. 2016 ; Breidenbach et al. 2025 . It is important to note that the boundaries themselves are relatively insensitive to these uncertainties since they are quite sharp. Taking all these factors into account, the comparison of S(\mathbf{q}) between DMRG and neutron scattering allows us to constrain the exchange parameters to the region \delta R\lesssim 1 as shown in Fig. 4 , although it does not pinpoint one specific parameter set. Despite this remaining uncertainty, the constrained region still lies entirely within the QSL phase determined by DMRG simulations. These comparisons provide strong additional evidence supporting the presence of the QSL state in both kagome materials. Figure 5: The dynamic spin structure factor S(\mathbf{q},\omega) , comparing results from DDMRG simulations (form factor adjusted) and measured inelastic neutron scattering data in the (HK0) plane for Zn-barlowite Breidenbach et al. 2025 for (a) \hbar\omega=0.15J and (b) \hbar\omega=0.4J at T=1.7 K. For the comparisons, the results are D_{6} -symmetrized and shown on relative intensity scales with the structural Brillouin zones overlaid. To further bridge the model with experiments, we have performed dynamical-DMRG (DDMRG) Jeckelmann 2002 ; Zhu et al. 2019 simulation and obtained the dynamic spin structure factor S(\mathbf{q},\omega) , which we directly compared with recent inelastic neutron scattering measurements on Zn barlowite Breidenbach et al. 2025 . The DDMRG simulations are carried out on XC8-11 cylinder with one additional edge column to reduce the boundary effect, using a broadening factor \eta=0.1J_{1} and keeping a maximum bond dimension of 2400. More details of DDMRG are provided in the End Matter. We choose two characteristic energy slices \hbar\omega=0.15J_{1} and \hbar\omega=0.4J_{1} , and compared the resulting S(\mathbf{q},\omega) between DDMRG simulation using ( J_{2}=-0.02J_{1} , D=0 ) and INS (using J_{1}=15 meV) in Fig. 5 . The agreement between DDMRG simulations (symmetrized and magnetic form factor adjusted) and neutron scattering data is excellent, especially for the higher energy slice ( \hbar\omega=0.4J_{1} ). The slight discrepancy near \rm K^{\prime} for \hbar\omega=0.15J_{1} may stem from finite-size effects of the XC8 cylinder that frustrates the \sqrt{3}\times\sqrt{3} order which contributes to the intensity at the \rm K^{\prime} point, as well as non-intrinsic scattering from interlayer impurities which exist at low energies Breidenbach et al. 2025 . We have also examined the pure Heisenberg model ( J_{1} only, see the End Matter), which shows a stronger peak at \rm\Gamma^{\prime} point with slightly poorer agreement with the neutron data. This suggests that a small ferromagnetic J_{2} is likely present in Zn-barlowite, consistent with a previous analysis Breidenbach et al. 2025 . A more precise determination of the parameters will require multiple simulations with different parameter sets (J_{2},D) , which we will explore in future work. Summary: We performed large-scale DMRG simulations of the kagome antiferromagnet, incorporating both the NNN exchange J_{2} and DM interaction D , a model of direct relevance to real kagome materials. Through extensive simulations on 2D-like lattice geometries, we have identified an extended QSL phase in the ground-state phase diagram and established its robustness across a broad parameter space of J_{2} and D . By comparing both static and dynamic spin structure factors from DMRG simulations with inelastic neutron scattering data, we further refined the exchange parameters for Zn-barlowite and herbertsmithite. Our findings reveal that the microscopic models for both materials reside within the QSL phase established by DMRG simulations, providing additional strong support for the presence of QSL ground states in these kagome compounds. Continued experimental advances in inelastic neutron scattering on single crystals, together with the development of large-scale DDMRG simulations incorporating nonzero J_{2} , D , and exchange anisotropy Han et al. 2012b terms, will enable a more precise determination of the model Hamiltonians for these QSL candidates, which is key to validating leading theoretical ideas. Data availability : The data used to generate the figures are deposited in doi.org/10.6084/m9.figshare.28755095 . Acknowledgment: SJ acknowledges helpful discussion with Steve White regarding DDMRG. ACC acknowledges Aaron Breidenbach for help with sample preparation of the Zn-barlowite. This work was supported by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division, under contract DE-AC02-76SF00515. This research used resources at the Spallation Neutron Source, Department of Energy (DOE), Office of Science User Facilities, operated by the Oak Ridge National Laboratory. DDMRG calculations were performed on resources of the National Energy Research Scientific Computing Center, supported by the U.S. Department of Energy under contract DEAC02-05CH11231. End Matter Line plot of \delta R(J_{2},D) beyond cutoff. – In the contour plots in Fig. 4 , we apply an upper cutoff of 1.5 on the color scale to emphasize the variation below this cutoff. Here in Fig. 6 we show the line plot of \delta R(J_{2},D) along D=0 and D=0.05 . As we can see, \delta R is relatively flat below the cutoff but increases rapidly above the cutoff. This justifies constraining the (J_{2},D) for the materials to the region \delta R(J_{2},D)\lesssim 1 , since the region outside has a clear deviation from experiments. Figure 6: Line plot of \delta R(J_{2},D) at fixed D=0 and D=0.05 between DMRG simulation and INS data for both materials. The black dashed line is the cutoff of 1.5 used in Fig. 4 . Dynamical DMRG simulation. – Our goal is to calculate the dynamic spin structural factor, defined as: \begin{split}S(\mathbf{q,\omega})=&\frac{-1}{\pi N}\sum_{\begin{subarray}{c}i,j=1\\ \alpha=x,y,z\end{subarray}}^{N}e^{i\mathbf{q}\cdot(\mathbf{r}_{i}-\mathbf{r}_{j})}\times\\ &\text{Im}\langle 0|S^{\alpha}_{\mathbf{r}_{i}}\frac{1}{E_{0}+\omega-H+i\eta}S^{\alpha}_{\mathbf{r}_{j}}|0\rangle,\end{split} (4) where |0\rangle is the ground state obtained by DMRG that has energy E_{0} , and \eta is the broadening factor. We define the correction vector |c^{\alpha}_{\mathbf{r}_{i}}(i,\omega)\rangle that satisfies: (E_{0}+\omega-H+i\eta)|c^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle=S^{\alpha}_{\mathbf{r}_{j}}|0\rangle, (5) the S(\mathbf{q,\omega}) can then be computed with: S(\mathbf{q,\omega})=\frac{-1}{\pi N}\sum_{\begin{subarray}{c}i,j=1\\ \alpha=x,y,z\end{subarray}}^{N}e^{i\mathbf{q}\cdot(\mathbf{r}_{i}-\mathbf{r}_{j})}\times\text{Im}\langle 0|S^{\alpha}_{\mathbf{r}_{i}}|c^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle. (6) The correction vector is separated into real and imaginary parts: |X^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle=\text{Re}|c^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle and |Y^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle=\text{Im}|c^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle , so that it can be solved with real arithmetic. Its imaginary part satisfies: [(E_{0}+\omega-H)^{2}+\eta^{2}]|Y^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle=-\eta S^{\alpha}_{\mathbf{r}_{j}}|0\rangle, (7) based on which the real part can also be computed: |X^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle=\frac{H-E_{0}-\omega}{\eta}|Y^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle. (8) The dynamical-DMRG algorithm Jeckelmann 2002 solves Eq. 7 using the same sweep scheme as the DMRG algorithm White 1992 ; White 1993 , replacing the original eigenvalue problem with the linear problem in Eq. 7 . At each step of the sweep, an effective superblock Hamiltonian H^{\rm eff} is formed by projecting the original Hamiltonian on the reduced basis. H^{\rm eff} approximates the H in Eq. 7 , and this linear equation is solved using the conjugate gradient method. To minimize the effect of this approximation, |X^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle needs to be state-averaged with |Y^{\alpha}_{\mathbf{r}_{j}}(\omega)\rangle in DDMRG. We implemented the DDMRG algorithm using the ITensor library Fishman et al. 2022 . In this paper, we select an \eta=0.1J_{1} as the broadening factor. We also compared our results with using a smaller \eta=0.05J_{1} in Fig. 7 and found only very minor quantitative changes (except for the overall scale that can change with \eta ), demonstrating that our results are robust with respect to the choice of \eta . Because the SU(2) symmetry is preserved in the QSL state we study, we replace the summation over spin component \alpha in Eq. 4 by the z component multiplied by three. To reduce boundary effects, the DDMRG simulations are carried out on XC8-11 cylinder with one additional edge column, and the Fourier transform in Eq. 4 uses spin correlations from the central 4\times 4 unit cells (of size XC8-4). Figure 7: The S(\mathbf{k},\omega) obtained by DDMRG with ( J_{2}=-0.02J_{1} , D=0 ) at \hbar\omega=0.2J_{1} (top) and \hbar\omega=0.4J_{1} (bottom), (a)(b) with \eta=0.1J_{1} and (c)(d) \eta=0.05J_{1} . The white dashed lines are the extended BZ boundaries. The results are D_{6} symmetrized and interpolated. Experiment and sample details . The reported neutron data were collected using the CNCS spectrometer at Oak Ridge National Laboratory’s Spallation Neutron Source. Spectra were measured at sample temperatures T=1.7 K and T=40 K and with incident neutron energies E_{i}=3.32 meV and E_{i}=12 meV over 360 ∘ scans in the HK0 scattering plane. The Zn-barlowite sample is made of 0.76 g of 190 co-aligned crystals Breidenbach et al. 2025 and the herbertsmithite sample is made of fifteen co-aligned crystals with a total mass of 1.2 g Han et al. 2016 . J_{2} comparison. — Here in Fig. 8 we compare the S(\mathbf{k},\omega) obtained by DDMRG with a ferromagnetic J_{2}=-0.02J_{1} and without J_{2} , both with D=0 . It is evident that the inclusion of a ferromagnetic J_{2} smears the peak at the \rm\Gamma^{\prime} point, leading to a closer resemblance with the INS data as shown in Fig. 5 . Figure 8: The S(\mathbf{k},\omega) obtained by DDMRG at \hbar\omega=0.2J_{1} (top) and \hbar\omega=0.4J_{1} (bottom), (a)(b) with J_{2}=-0.02J_{1} and (c)(d) without J_{2} . D=0 for all calculations. The white dashed lines are the extended BZ boundaries. The intensity is normalized to [0 1] in each figure. The results are D_{6} symmetrized.