This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Magnetic field-driven transition between valence bond solid and antiferromagnetic order in distorted triangular lattice

Yasuhiro Shimizu Department of Physics, Nagoya University, Chikusa, Nagoya 464-8602, Japan.    Mitsuhiko Maesato Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan.    Makoto Yoshida Institute for Solid State Physics, University of Tokyo, Kashiwa, Chiba 277-8581, Japan.    Masashi Takigawa Institute for Solid State Physics, University of Tokyo, Kashiwa, Chiba 277-8581, Japan.    Masayuki Itoh Department of Physics, Nagoya University, Chikusa, Nagoya 464-8602, Japan.    Akihiro Otsuka Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan. Research Center for Low Temperature and Materials Sciences, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan    Hideki Yamochi Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan. Research Center for Low Temperature and Materials Sciences, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan    Yukihiro Yoshida Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan. Faculty of Agriculture, Meijo University, Tempaku-ku, Nagoya 468-8502, Japan.    Genta Kawaguchi Division of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan.    David Graf National High Magnetic Field Laboratory, Florida State University, Tallahassee, 32310 FL.    Gunzi Saito Faculty of Agriculture, Meijo University, Tempaku-ku, Nagoya 468-8502, Japan. Toyota Physical and Chemical Research Institute, Nagakute, Aichi 480-1192, Japan.
Abstract

A molecular Mott insulator κ\kappa-(ET)2B(CN)4 [ET = bis(ethylenedithio)tetrathiafulvalene] with a distorted triangular lattice exhibits a quantum disordered state with gapped spin excitation in the ground state. 13C nuclear magnetic resonance, magnetization, and magnetic torque measurements reveal that magnetic field suppresses valence bond order and induces long-range magnetic order above a critical field 8\sim 8 T. The nuclear spin-lattice relaxation rate 1/T11/T_{1} shows persistent evolution of antiferromagnetic correlation above the transition temperature, highlighting a quantum spin liquid state with fractional excitations. The field-induced transition as observed in the spin-Peierls phase suggests that the valence bond order transition is driven through renormalized one-dimensionality and spin-lattice coupling.

preprint: APS/123-QED

I Introduction

Quantum magnets featured by macroscopic quantum entanglements host exotic quasiparticles and topological order [1, 2, 3]. In frustrated spin systems, a valence bond order state with translational symmetry breaking competes with the magnetic order and quantum spin liquid states [4, 5, 6, 7]. The phase transition between magnetic order and valence bond crystal phases may occur through continuous quantum phase transition [3, 8], instead of the first-order transition that separates the two phases having competing order parameters in Landau-Ginzburg-Wilson paradigm. The realization of deconfined quantum criticality with fractional excitation has been a fundamental issue of many-body quantum physics [4, 9, 10, 11].

Under an intense magnetic field, the valence bond order state is destabilized and exhibits a transition into long-range order states [12, 13]. The quasiparticle in the spin-dimer state is a spin S=1S=1 triplon localized on the lattice site. The field-induced transition involving long-range order of transverse magnetization is regarded as Bose-Einstein condensation of triplons, as observed in several quantum magnets [14, 15, 16, 17, 18, 19, 20, 12, 13]. On one-dimensional (1D) systems, the spin-Peierls transition occurs through spin-lattice coupling and is suppressed by forming a soliton lattice under high magnetic field [21, 22, 23, 24]. By contrast the spin-singlet formation in 2D frustrated spin systems can be highly degenerate and involve fractionalized quasiparticles [9].

Molecular Mott insulators with a triangular lattice such as κ\kappa-(ET)X2{}_{2}X and XX[Pd(dmit)2]2 (XX: counter ions) may provide the candidate showing the deconfined criticality. Depending on the strength and anisotropy of intermolecular interactions, a series of materials exhibit fertile ground states including quantum spin liquid [25, 26, 27, 28, 29, 30], antiferromagnetic order [31, 32], and valence bond order [33, 34]. Some of them display superconductivity under pressure [35, 36]. Here we focus on a Mott insulator κ\kappa-(ET)2B(CN)4 having an anisotropic triangular lattice [33]. The anisotropy of exchange interactions is evaluated as J/J=2J^{\prime}/J=2 (J=118J=118 K, J=236J^{\prime}=236 K) from the spin susceptibility [37], where JJ (or transfer integral tt) and JJ^{\prime} (tt^{\prime}) form a square lattice and a chain, respectively [Fig. 1(a)]. The anisotropy increases up to J/J3J^{\prime}/J\sim 3, as temperature (TT) decreases. It agrees with the calculated transfer anisotropy t/t=1.4t^{\prime}/t=1.4 or (t/t)2(t^{\prime}/t)^{2} = 2 [33] and the optical measurement [38]. Under pressure up to 2.5 GPa, the system remains Mott insulating. Thus the material is located far from the Mott transition at ambient pressure and provides an ideal Heisenberg antiferromagnet with anisotropic triangular lattice.

The spin susceptibility of κ\kappa-(ET)2B(CN)4 exponentially decreases below Tc5T_{c}\sim 5 K without showing an indication of long-range magnetic order in the 1H NMR spectrum [33]. Thus the ground state likely accompanies the valence bond order with translational symmetry breaking or the spin-Peierls transition, whereas the crystal structure has not been solved below TcT_{c}. The 1H NMR nuclear-lattice relaxation rate T11T_{1}^{-1} exhibits weak TT dependence above TcT_{c}, consistent with the behavior of the low-dimensional quantum antiferromagnet in the quantum critical region [39]. Considering the low TcT_{c}, an application of magnetic field is expected to induce quantum phase transition, which gives an insight into the origin of the valence bond order and the quasiparticle property.

Refer to caption
Figure 1: (a) Anisotropic triangular lattice of molecular dimers in κ\kappa-(ET)2B(CN)4 with two kinds of transfers tt and tt^{\prime} above TcT_{c}. Spin-1/2 on each ET dimer is paired to form a valence bond crystal below Tc=5T_{c}=5 K. (b) Magnetic phase diagram based on NMR (closed symbols) and spin susceptibility χ\chi (open symbols). Circles and diamonds with the uncertainty denote the valence bond order and antiferromagnetic transition temperatures, respectively, which are obtained from the temperature dependence of nuclear spin-lattice relaxation rate T11T_{1}^{-1}.

In this Article, we investigate the field-driven transition from the valence bond order phase in κ\kappa-(ET)2B(CN)4. The local field distribution and spin excitation are probed by 13C NMR spectroscopy down to low temperatures under magnetic field up to 14 T. Together with the magnetization and torque measurements, we uncover the ground state and low-energy properties across the field-driven transition. The magnetic field (HH) versus TT phase diagram based on the experimental results [Fig. 1(b)] is compared with the quantum spin liquid and spin-Peierls systems.

II Experimental

Single crystals of κ\kappa-(ET)2B(CN)4 were grown by an electrooxidation method [33]. The magnetization was measured with a superconducting quantum interference device under 0.5 – 7 T. The high-field magnetic torque for a single crystal was measured using a piezoresistive cantilever at the National High Magnetic Field Laboratory in Tallahassee. The magnetic field direction was tilted by 16 from the acac plane. For NMR experiments, the 90% 13C isotope was selectively enriched to the central double-bonded carbon sites of ET [Fig. 1(a)] [40]. NMR measurements were conducted on a single crystal under the external magnetic field oriented along the aa axis (<1<1^{\circ}) of the conducting plane by utilizing a two-axis goniometer. The frequency-swept NMR spectrum was obtained from the spin echo measurement with the pulse width tπ/2=t_{\pi/2}= 0.5 μ\mus and tπ=1t_{\pi}=1 μ\mus, and the sum was taken by 0.2 MHz steps covering the whole spectrum at low temperatures [Fig. 5(b)]. The nuclear magnetization recovery after the saturation followed a single exponential function above 5 K and becomes a stretched exponential function M(t)/M0=1exp[(t/T1)β]M(t)/M_{0}=1-{\rm exp}[-(t/T_{1})^{\beta}] with the exponent β=0.31\beta=0.3-1 below 5 K, which is nearly independent of HH.

III Experimental Results

III.1 Spin susceptibility and Knight shift

Refer to caption
Figure 2: (a) Spin susceptibility χ\chi for a polycrystalline sample of κ\kappa-(ET)2B(CN)4 at 1 T [33] after subtracting the core diamagnetic susceptibility (χdia=4.52×104\chi_{\rm dia}=-4.52\times 10^{-4} emu mol-1). The red curve is a fitting result by a high-temperature series expansion [37]. Inset: Temperature dependence of χ\chi at low temperatures under the magnetic field of 0.5, 1, 3, and 5 T. (b) Inverse temperature dependence of χ\chi for the evaluation of spin gap with the exponential function χexp(Δ/T)\chi\sim{\rm exp}(-\Delta/T).

The spin susceptibility χ\chi of κ\kappa-(ET)2B(CN)4 exhibits a broad maximum around 130 K [33], characteristic to the low-dimensional antiferromagnet, as shown in Fig. 2(a). The TT dependence is fitted into a high-TT series expansion of the Heisenberg model with anisotropic triangular lattice (J=118J=118 K, J=236J^{\prime}=236 K) [37]. Below TcT_{c} = 5 K, χ\chi measured at 1 T decreases continuously, indicating the second-order phase transition. The spin gap was evaluated as Δ\Delta = 27 K for the singlet-triplet model [33] at 1 T. Here we employ the simple exponential function [Fig. 2(b)], yielding Δ\Delta = 16, 14.8, and 11.1 K under HH = 1, 3, and 5 T, respectively. The negligible Curie tail increase in χ\chi at low temperatures certificates the high crystal quality.

Refer to caption
Figure 3: (a) 13C NMR spectra at HH = 3.9 T along the aa axis in κ\kappa-(ET)2B(CN)4. Lorentzian fitting curves are shown at the lowest temperature. (b) Knight shifts, K(1)K(1) and K(2)K(2), for two carbon sites, C(1) and C(2), which are measured from the reference (tetramethylsilane). A green curve denotes the spin susceptibility χ\chi at 1.0 T (the right-hand axis) [33].

The 13C NMR spectrum provides the local spin susceptibility and the local field distribution at nuclear sites through the hyperfine interaction. In κ\kappa-(ET)2B(CN)4, each ET molecule has two 13C sites, all of which in the unit cell (orthorhombic PnmaPnma) are equivalent [Fig. 1(a)]. Therefore, the NMR spectrum in Fig. 3(a) consists of two doublets due to dipole-coupled two 13C spins [41]. Upon cooling, the spectrum gradually broadens toward TcT_{c}, coinciding with the intensity reduction due to fluctuations and the inhomogneous field due to the generation of paramagnetic domains around the structural transition. Below TcT_{c}, the spectrum turns to be narrow and remains only the doublet splitting. The splitting frequency is about 3/23/2 times larger than that observed at high temperatures, which represents a crossover from the unlike-spin to like-spin coupling, as the spin susceptibility vanishes below TcT_{c} [41]. Since 13C spins (I=1/2I=1/2) do not have electric quadrupole interaction, structural fluctuations is observed through a small nuclear dipole coupling and paramagnetic domains.

The residual broad tails of the spectrum at low temperatures can be attributed to staggered moments nucleated around the domains through the Dzyaloshinkii-Moriya (DM) interaction [32]. The sharp central line means that the majority of spins remain nonmagnetic.

The Knight shift KK proportional to the local spin susceptibility χloc\chi_{\rm loc} is defined for two carbon sites on each ET molecule, termed as K(1)K(1) and K(2)K(2). As shown in Fig. 3(b), K(1)K(1) exhibits a broad maximum around 130 K as observed in χ\chi, reflecting the short-range antiferromagnetic correlation. In contrast, K(2)K(2) monotonously decreases upon cooling. The site dependence implies that the hyperfine coupling and the chemical shift vary with TT due to the anisotropic lattice contraction [33].

Refer to caption
Figure 4: 13C Knight shifts plotted against χ\chi for 10 – 50 K. The linearity yield the hyperfine coupling constant for two 13C sites. The yy-intercept gives the chemical shift.

At low temperatures below 50 K, the Knight shifts well scale to χ\chi. The linearity in the KχK-\chi plot for 105010-50 K yields the hyperfine coupling constant HhfH_{\rm hf}, as shown in Fig. 4. In the present field direction (HaH\parallel a), we obtained HhfH_{\rm hf} = 0.34(2) T/μB\mu_{\rm B} for C(1) and 0.18(4) T/μB\mu_{\rm B} for C(2). The difference reflects the asymmetric spin density distribution due to the molecular dimerization [32]. Here HhfH_{\rm hf} is composed of the Fermi contact and dipolar interaction terms. The anisotropic dipole term may depend on temperature when the principal axis governed by the strength of the dimerization changes upon cooling. Similar behavior has been observed in the other members of κ\kappa-(ET)2X [32, 42].

III.2 Magnetic field driven transition

Refer to caption
Figure 5: Magnetic torque τ\tau divided by magnetic field HH at low temperatures. The data at each temperature include a constant offset for clarity.

The field sensitivity of the spin gap implies that a quantum phase transition occurs at the critical field HcH_{c} where Δ\Delta goes to zero or competes with the other order parameter. For investigating the field-driven transition, we measured the magnetic torque τ\tau responsive to the transverse magnetization, τ/HMT\tau/H\sim M_{T}, as shown in Fig. 5. In a paramagnetic state above TcT_{c}, τ/H\tau/H is governed by the gg-value anisotropy and increases with the external field HH. The slop is nearly independent of temperature at high fields.

Below 4.5 K, τ/H\tau/H becomes nonlinear to HH. τ/H\tau/H behaves constantly at low fields, reflecting χ0\chi\simeq 0 with a negligible Curie impurity component. Then τ/H\tau/H starts to increase around 8 T, which clearly indicates a magnetic transition from the nonmagnetic state into the magnetically ordered state or the paramagnetic spin liquid state. Below 2.5 K, a maximum appears in τ/H\tau/H around 11–12 T. After showing a minimum around 24–28 T, τ/H\tau/H linearly increases with HH without saturation up to 35 T. However, the nature of the strong field-induced phase is unclear at the present stage.

Refer to caption
Figure 6: (a) Magnetic field dependence of the 13C NMR spectrum for κ\kappa-(ET)2B(CN)4 at 1.4 K. Frequency ω\omega was converted to the local field ΔB\Delta B using a relation ΔB=Δω/γ0\Delta B=\Delta\omega/\gamma_{0} (γ0=10.7054\gamma_{0}=10.7054 MHz/T). (b) Definition of the 1/2 and 1/10 linewidths for TT = 1.4 K and HH = 12 T, where the spectrum is constructed as a sum of Fourier transformed spin-each intensities taken with a 0.2 MHz step. (c) Magnetic field dependence of the 1/2 and 1/10 linewidths.

The magnetic field dependence of 13C NMR spectrum was measured under the magnetic field of 3.9 – 14 T at 1.4 K, as displayed in Fig. 6(a). The spectrum gradually spreads even at low fields, indicating the evolution of the local field distribution upon increasing HH. Above 8.5 T, the spectral broadening becomes prominent at low temperatures. The NMR spectrum eventually changes into a two-step bell shape. The emergence of the huge local field signals long-range magnetic order. Under HH = 14 T, the maximum local field reaches ±0.04\pm 0.04 T at the tails of the spectrum.

As shown in Fig. 6(b,c), the linewidths defined at the 1/2 and 1/10 of the maximum intensity evolve above the critical field HcH_{c}\sim 8 T and levels off at high fields, consistent with the magnetic torque result. The 1/2 and 1/10 widths measure the evolution of the order parameter at two 13C sites with the different HhfH_{\rm hf}. Indeed, the 1/2 width (0.03 T) and 1/10 width (0.07 T) approximately scale to HhfH_{\rm hf} of two 13C sites. The result yields the maximum moment of \sim0.1 μB\mu_{\rm B}, which continuously grows above HcH_{c} and saturates at high fields.

Refer to caption
Figure 7: Temperature dependence of the 13C NMR spectrum for the single crystal of κ\kappa-(ET)2B(CN)4 under the magnetic field of (a) 8.5 and (b) 14 T along the aa axis.

The temperature dependence of NMR spectra above HcH_{c} is shown in Fig. 7. At 8.5 T, the spectral broadening starts below 10 K and continues upon cooling, while leaving a narrow line at the central position [Fig. 7(a)]. It suggests the coexistence of the magnetic phase and the valence bond order phase, or the formation of soliton lattice as observed in the 1D chains [22]. However, the singular edge peak characteristic to solitons was not detected within the experimental uncertainty. At 14 T, the site-dependent broadening becomes clearer above TNT_{\rm N}. The asymmetric spectral shape in the paramagnetic state becomes nearly symmetric below 2.5 K, which gives evidence for the magnetic order transition.

In reference to the spin structure of κ\kappa-(ET)2Cu[N(CN)2]Cl belonging to the same space group as κ\kappa-(ET)2B(CN)4, the colinear magnetic order splits the 13C NMR spectrum under the magnetic field along the easy axis (HcH\parallel c) and gives only a positive shift for the perpendicular direction (H||aH||a) [43]. The observed continuous and symmetric broadening points to an incommensurate magnetic structure such as a spiral order expected for a triangular lattice antiferromagnet [7, 44, 45]. The full determination of the spin structure requires the HhfH_{\rm hf} tensor and the angular dependence measurements. The sizable moment contraction highlights a crucial role of quantum fluctuations for the system nearby the disordered state.

III.3 Spin dynamics across field driven transition

Refer to caption
Figure 8: (a) Temperature dependence of the 13C nuclear spin-lattice relaxation rate T11T_{1}^{-1} at 3.9–7.8 T. 1H NMR T11T_{1}^{-1} (a long component) at 2.0 T scale to 13C data by multiplying 10210^{2} [33]. TcT_{c} defined by the onset of T11T_{1}^{-1} drop is marked by arrows. (b) Inverse temperature dependence of T11T_{1}^{-1} for evaluating the spin gap Δ\Delta. (c) Field dependence of Δ\Delta (the right-hand axis) and the contour plot of T11T_{1}^{-1} obtained from the temperature and field dependence measurement. Solid and open circles are evaluated from T11T_{1}^{-1} and χ\chi, respectively.

The spin excitation is investigated by the nuclear spin-lattice relaxation rate T11T_{1}^{-1} across HcH_{c}. As shown in Fig. 8, T11T_{1}^{-1} is nearly independent of TT above TcT_{c}, as expected in low-dimensional quantum antiferromagnets for a TT range lower than the exchange coupling (T<JT<J) [39]. The behavior differs from that observed in a triangular lattice system having the spin liquid ground state, as discussed below. T11T_{1}^{-1} decreases steeply as the spin gap Δ\Delta opens below TcT_{c} in a low field range below 8 T. T11T_{1}^{-1} is enhanced with increasing HH, as TcT_{c} and Δ\Delta are suppressed toward Hc8H_{c}\sim 8 T. The obtained field dependence of TcT_{c} is plotted in Fig. 1(b), consistent with the result of χ\chi.

We evaluated Δ\Delta by fitting T11T_{1}^{-1} into an exponential function \sim exp(Δ/T)(-\Delta/T) [Fig. 8(b)]. We obtained Δ=14.5,12.0,7.2\Delta=14.5,12.0,7.2 K at HH = 3.9, 5.0, 6.9 T, respectively, and plotted in Fig. 8(c). The results are consistent with those obtained from χ(T)\chi(T) [Fig. 2(b)]. Since T11T_{1}^{-1} measures the dynamical spin correlation summed over the wave vector space, the result represents the presence of the spin excitation gap independent of the wave vector. We find that Δ\Delta linearly decreases with increasing HH, Δ(HcH)z\Delta\sim(H_{c}-H)^{z}, with Hc8.2(2)H_{c}\sim 8.2(2) T and the critical exponent z=1.0(3)z=1.0(3). The HH and TT dependence of T11T_{1}^{-1} is also illustrated as a contour plot [Fig. 4(c)]. Approaching HcH_{c}, the TT independent region persists to low temperatures, as expected in the gapless spinon regime.

Refer to caption
Figure 9: (a) Temperature dependence of T11T_{1}^{-1} above 7.8 T. Arrows point to the magnetic order transition at TNT_{\rm N}. (b) Linewidth of 13C NMR spectrum, defined by 1/10 of the peak intensity in Fig. 6(b).

Above HcH_{c}, T11T_{1}^{-1} exhibits a peak structure at 22.52-2.5 K due to the slowing-down of spin fluctuations, as shown in Fig. 9(a). The peak determines the magnetic order transition temperature TNT_{\rm N}. It is nearly independent of HH, as plotted in Fig. 1(b). Below TNT_{\rm N} one can expect an increase in the linewidth of the 13C NMR spectrum, which measures the evolution of the order parameter. As shown in Fig. 9(b), the linewidth remains small at 7.8 T just below HcH_{c}. The spectral broadening starts from T>TNT>T_{\rm N} above 8.5 T due to fluctuations and inhomogeneity, while the broadening levels off below TNT_{\rm N} defined by T11T_{1}^{-1}.

IV Discussion

In this section, we focus on the nature of the dynamical spin correlation in the paramagnetic spin liquid phase. We also discuss the origin for the valence bond order transition based on the magnetic phase diagram as well as the possible magnetic phase at high fields.

The series of κ\kappa-(ET)X2{}_{2}X [XX = Cu2(CN)3, Ag2(CN)3] have been extensively investigated as frustrated quantum antiferromagnets with triangular lattice close to Mott transition [25, 32, 30], where the anisotropy t/tt^{\prime}/t is close to or smaller than unity. κ\kappa-(ET)2B(CN)4 is the rare example of the frustrated triangular lattice with one-dimensional anisotropy (t/t>1t^{\prime}/t>1). Thus it is interesting to compare the difference in the dynamical spin correlation between these systems. In the temperature range lower than the antiferromagnetic exchange interaction, the system is regarded as quantum spin liquid involving low-lying excitations. T11T_{1}^{-1} measures the dynamical spin susceptibility in the low energy limit [39, 9]. On the isotropic triangular lattice such as κ\kappa-(ET)2Ag2(CN)3 showing a quantum spin liquid state, T11T_{1}^{-1} follows a power law Tη\sim T^{\eta} (20<T<20020<T<200 K) with the anomalous exponent η=0.40.5\eta=0.4-0.5, as shown in Fig. 10 [32, 42]. Interestingly, the exponent is close to that expected for deconfined spinon excitations (η=0.6\eta=0.6) in the frustrated model [9].

Refer to caption
Figure 10: Temperature dependence of T11T_{1}^{-1} for 13C NMR in κ\kappa-(ET)X2{}_{2}X, XX = B(CN)4, Cu2(CN)3 [42], and Ag2(CN)3 [32].

In contrast, T11T_{1}^{-1} for κ\kappa-(ET)2B(CN)4 follows a power law Tη\sim T^{\eta} with the small exponent η=0.19(1)\eta=0.19(1) in a wide range (5<T<1005<T<100 K), as shown in Fig. 10. Here the strong frustration effect is manifested in the low TcT_{c} and TNT_{\rm N} compared with the exchange coupling. The exponent is rather close to that of the 1D quantum critical antiferromagnet, where T11T_{1}^{-1} behaves TT-independent (Tη\sim T^{\eta}, η=0\eta=0), as the antiferromagnetic correlation length ξ\xi diverges with ξTz\xi\sim T^{-z} (the dynamical exponent z=1z=1) at the specific wave vector [39]. The behavior was observed in Sr2CuO3 [46] and (TMTTF)2PF6 [47].

On the magnetic phase diagram [Fig.1(b)], the valence bond order phase is suppressed toward the critical field Hc8H_{c}\sim 8 T, as observed by magnetic torque and NMR spectrum measurements. It is also consistent with the field dependence of the spin gap Δ\Delta obtained from T11T_{1}^{-1}. The continuous suppression of Δ\Delta implies the quantum phase transition in the frustrated spin system [3, 11, 10]. However, the incommensurate magnetic order phase appears from the finite temperature across HcH_{c}. It implies that the second order valence bond order transition is terminated by the first-order magnetic transition at low temperatures as anticipated from the mean-field theory [48, 49].

The shape of phase diagram is similar to those of the quasi-1D spin-Peierls system such as CuGeO3 and (TMTTF)2PF6[50, 22, 47, 51]. It suggests that the valence bond order occurs through the lattice dimerization along the cc axis. Indeed, the crystal structure becomes more 1D upon cooling [33]. This tendency is distinct from that observed in the other κ\kappa-(ET)X2{}_{2}X members with t/t<1t^{\prime}/t<1 [30]. Thus the renormalized 1D spin correlation can develop in the gapless spin liquid regime above TcT_{c} [52] and eventually couple the lattice degrees of freedom. Furthermore, the discrete (not polymeric) counter ion and the absence of the structural disorder allows the spin dimerization along one direction.

Focusing on the magnetically ordered phase at high fields, the magnetic structure would be highly degenerate and nontrivial on the triangular lattice. Theoretical calculations have been reported based on the Hubbard and Heisenberg model as a function of anisotropy of triangular lattice, t/tt^{\prime}/t or J/JJ^{\prime}/J [53, 7, 45]. For J/J>1J^{\prime}/J>1, the magnetic ground state of Mott insulator can be both quantum spin liquid and antiferromagnetic orders such as colinear and spiral spin structures. Indeed, the isostructure Mott insulator κ\kappa-(ET)2CF3SO3 (t/t=1.8t^{\prime}/t=1.8) exhibits the magnetically ordered ground state with an incommensurate structure below TN=2.5T_{\rm N}=2.5 K [54]. Therefore, the present system is naturally expected to exhibit the long-range magnetic order, as the valence bond crystal is suppressed by the magnetic field.

V Conclusion

In conclusion, the field-induced quantum phase transition was observed from the valence bond order to the antiferromagnetic long-range order state by 13C NMR and magnetization measurements in the Mott insulator κ\kappa-(ET)2B(CN)4 having a distorted triangular lattice. We found that the spin excitation gap is linearly suppressed by magnetic field toward the critical point. As the valence bond order phase melts, the quantum spin liquid with gapless excitation appears above the critical field. The obtained phase diagram manifests the competing magnetic ground states in the frustrated magnet. The significant role of quantum fluctuations are highlighted as the reduction of the antiferromagnetic moment and the persistent low-lying excitation down to low temperatures.

Acknowledgements

This work was financially supported by Grants-in-Aid for Scientific Research (Grants No. JP16H02206, JP16H04012, JP17H05151, JP18H04223, JP19H05824, JP19H01837, and JP20H02709) from JSPS, and partly supported by the Kyoto University Foundation.

References

  • Anderson [1973] P. W. Anderson, Resonating valence bonds: a new kind of insulator?, Mater. Res. Bull. 8, 153 (1973).
  • Wen [2002] X.-G. Wen, Quantum orders and symmetric spin liquids, Phys. Rev. B 65, 165113 (2002).
  • Senthil et al. [2004a] T. Senthil, A. Vishwanath, L. Balents, S. Sachdev, and M. P. A. Fisher, Deconfined quantum critical points, Science 303, 1490 (2004a).
  • Read and Sachdev [1991] N. Read and S. Sachdev, Large-N expansion for frustrated quantum antiferromagnets, Phys. Rev. Lett. 66, 1773 (1991).
  • Alicea et al. [2005] J. Alicea, O. I. Motrunich, M. Hermele, and M. P. A. Fisher, Criticality in quantum triangular antiferromagnets via fermionized vortices, Phys. Rev. B 72, 064407 (2005).
  • Hauke [2013] P. Hauke, Quantum disorder in the spatially completely anisotropic triangular lattice, Phys. Rev. B 87, 014415 (2013).
  • Tocchio et al. [2014] L. F. Tocchio, C. Gros, R. Valentí, and F. Becca, One-dimensional spin liquid, collinear, and spiral phases from uncoupled chains to the triangular lattice, Phys. Rev. B 89, 235107 (2014).
  • Sachdev [2011] S. Sachdev, Quantum Phase Transitions, 2nd ed. (Cambridge Univ. Press, London, 2011).
  • Senthil et al. [2004b] T. Senthil, L. Balents, S. Sachdev, A. Vishwanath, and M. P. A. Fisher, Quantum criticality beyond the Landau-Ginzburg-Wilson paradigm, Phys. Rev. B 70, 144407 (2004b).
  • Jian et al. [2018] C.-M. Jian, A. Thomson, A. Rasmussen, Z. Bi, and C. Xu, Deconfined quantum critical point on the triangular lattice, Phys. Rev. B 97, 195115 (2018).
  • Scammell and Sushkov [2015] H. D. Scammell and O. P. Sushkov, Violation of the spin-statistics theorem and the Bose-Einstein condensation of particles with half-integer spin, Phys. Rev. Lett. 114, 055702 (2015).
  • Zapf et al. [2014] V. Zapf, M. Jaime, and C. D. Batista, Bose-Einstein condensation in quantum magnets, Rev. Mod. Phys. 86, 563 (2014).
  • Giamarchi et al. [2008] T. Giamarchi, C. Ruegg, and O. Tchernyshyov, Bose-Einstein condensation in magnetic insulators, Nat. Phys. 4, 198 (2008).
  • Shirasawa et al. [2017] K. Shirasawa, N. Kurita, and H. Tanaka, Universality of magnetic-field-induced Bose-Einstein condensation of magnons, Phys. Rev. B 96, 144404 (2017).
  • Yamada et al. [2008] F. Yamada, T. Ono, H. Tanaka, G. Misguich, M. Oshikawa, and T. Sakakibara, Magnetic-field induced bose-einstein condensation of magnons and critical behavior in interacting spin dimer system TlCuCl3J. Phys. Soc. Jpn. 77, 013701 (2008).
  • Vyaselev et al. [2004] O. Vyaselev, M. Takigawa, A. Vasiliev, A. Oosawa, and H. Tanaka, Field-induced magnetic order and simultaneous lattice deformation in TlCuCl3Phys. Rev. Lett. 92, 207202 (2004).
  • Mazurenko et al. [2014] V. V. Mazurenko, M. V. Valentyuk, R. Stern, and A. A. Tsirlin, Nonfrustrated interlayer order and its relevance to the Bose-Einstein condensation of magnons in BaCuSi2O6Phys. Rev. Lett. 112, 107202 (2014).
  • Sebastian et al. [2006] S. E. Sebastian, N. Harrison, C. D. Batista, L. Balicas, M. Jaime, P. A. Sharma, N. Kawashima, and I. R. Fisher, Dimensional reduction at a quantum critical point, Nature (London) 441, 617 (2006).
  • Krämer et al. [2007] S. Krämer, R. Stern, M. Horvatić, C. Berthier, T. Kimura, and I. R. Fisher, Nuclear magnetic resonance evidence for a strong modulation of the Bose-Einstein condensate in BaCuSi2O6Phys. Rev. B 76, 100406 (2007).
  • Lancaster et al. [2014] T. Lancaster, P. A. Goddard, S. J. Blundell, F. R. Foronda, S. Ghannadzadeh, J. S. Möller, P. J. Baker, F. L. Pratt, C. Baines, L. Huang, J. Wosnitza, R. D. McDonald, K. A. Modic, J. Singleton, C. V. Topping, T. A. W. Beale, F. Xiao, J. A. Schlueter, A. M. Barton, R. D. Cabrera, K. E. Carreiro, H. E. Tran, and J. L. Manson, Controlling magnetic order and quantum disorder in molecule-based magnets, Phys. Rev. Lett. 112, 207201 (2014).
  • Hayashi and Ogata [2007a] Y. Hayashi and M. Ogata, Possibility of gapless spin liquid state by one-dimensionalization, J. Phys. Soc. Jpn. 76, 053705 (2007a).
  • Horvatic et al. [1999] M. Horvatic, Y. Fagot-Revurat, C. Berthier, G. Dhalenne, and A. Revcolevschi, NMR imaging of the soliton lattice profile in the spin-peierls compound CuGeO3, Phys. Rev. Lett. 83, 420 (1999).
  • Casola et al. [2013] F. Casola, T. Shiroka, A. Feiguin, S. Wang, M. S. Grbić, M. Horvatić, S. Krämer, S. Mukhopadhyay, K. Conder, C. Berthier, H.-R. Ott, H. M. Rønnow, C. Rüegg, and J. Mesot, Field-induced quantum soliton lattice in a frustrated two-leg spin-1/21/2 ladder, Phys. Rev. Lett. 110, 187201 (2013).
  • Kiryukhin et al. [1995] V. Kiryukhin, B. Keimer, and D. E. Moncton, Direct observation of a magnetic field induced commensurate-incommensurate transition in a spin-Peierls system, Phys. Rev. Lett. 74, 1669 (1995).
  • Shimizu et al. [2003] Y. Shimizu, K. Miyagawa, K. Kanoda, M. Maesato, and G. Saito, Spin liquid state in an organic Mott insulator with a triangular lattice, Phys. Rev. Lett. 91, 107001 (2003).
  • Itou et al. [2010] T. Itou, A. Oyamada, S. Maegawa, and R. Kato, Instability of a quantum spin liquid in an organic triangular-lattice antiferromagnet, Nat. Phys. 6, 673 (2010).
  • Isono et al. [2014] T. Isono, H. Kamo, A. Ueda, K. Takahashi, M. Kimata, H. Tajima, S. Tsuchiya, T. Terashima, S. Uji, and H. Mori, Gapless quantum spin liquid in an organic spin-1/2 triangular-lattice H3(Cat-EDT-TTF)2, Phys. Rev. Lett. 112, 177201 (2014).
  • Pratt et al. [2011] F. Pratt, P. J. Baker, S. J. Blundell, T. Lancaster, S. Ohira-Kawamura, C. Baines, Y. Shimizu, K. Kanoda, I. Watanabe, and G. Saito, Magnetic and nonmagnetic phases of a quantum spin liquid, Nature(London) 471, 612 (2011).
  • Hiramatsu et al. [2017] T. Hiramatsu, Y. Yoshida, G. Saito, A. Otsuka, H. Yamochi, M. Maesato, Y. Shimizu, H. Ito, Y. Nakamura, H. Kishida, M. Watanabe, and R. Kumai, Design and preparation of a quantum spin liquid candidate κ\kappa-(ET)2Ag2(CN)3 having a nearby superconductivity, Bull. Chem. Soc. Jpn. 90, 1073 (2017).
  • Kanoda and Kato [2011] K. Kanoda and R. Kato, Mott physics in organic conductors with triangular lattices, Ann. Rev. Conden. Matter Phys. 2, 167 (2011).
  • Miyagawa et al. [2004] K. Miyagawa, K. Kanoda, and A. Kawamoto, NMR studies on two-dimensional molecular conductors and superconductors: Mott transition in κ\kappa-(BEDT-TTF)2X, Chem. Rev. 104, 5635 (2004).
  • Shimizu et al. [2016] Y. Shimizu, T. Hiramatsu, M. Maesato, A. Otsuka, H. Yamochi, A. Ono, M. Itoh, M. Yoshida, M. Takigawa, Y. Yoshida, and G. Saito, Pressure-tuned exchange coupling of a quantum spin liquid in the molecular triangular lattice κ\kappa-(ET)2Ag2(CN)3Phys. Rev. Lett. 117, 107203 (2016).
  • Yoshida et al. [2015] Y. Yoshida, H. Ito, M. Maesato, Y. Shimizu, H. Hayama, T. Hiramatsu, Y. Nakamura, H. Kishida, T. Koretsune, C. Hotta, and G. Saito, Spin-disordered quantum phases in a quasi-one-dimensional triangular lattice, Nat. Phys. 11, 679 (2015).
  • Tamura et al. [2006] M. Tamura, A. Nakao, and R. Kato, Frustration-induced valence-bond ordering in a new quantum triangular antiferromagnet based on [Pd(dmit)2], J. Phys. Soc. Jpn. 75, 093701 (2006).
  • Shimizu et al. [2007] Y. Shimizu, H. Akimoto, H. Tsujii, A. Tajima, and R. Kato, Mott transition in a valence-bond solid insulator with a triangular lattice, Phys. Rev. Lett. 99, 256403 (2007).
  • Kurosaki et al. [2005] Y. Kurosaki, Y. Shimizu, K. Miyagawa, K. Kanoda, and G. Saito, Mott transition from a spin liquid to a fermi liquid in the spin-frustrated organic conductor κ\kappa-(ET)2Cu2(CN)3Phys. Rev. Lett. 95, 177001 (2005).
  • Zheng et al. [2005] W. Zheng, R. R. P. Singh, R. H. McKenzie, and R. Coldea, Temperature dependence of the magnetic susceptibility for triangular-lattice antiferromagnets with spatially anisotropic exchange constants, Phys. Rev. B 71, 134422 (2005).
  • Mizukoshi et al. [2018] K. Mizukoshi, Y. Nakamura, Y. Yoshida, G. Saito, and H. Kishida, Optical evaluation of electronic anisotropy in a triangular lattice system κ\kappa-(BEDT-TTF)2B(CN)4J. Phys. Soc. Jpn. 87, 104708 (2018).
  • Chubukov et al. [1994] A. V. Chubukov, S. Sachdev, and J. Ye, Theory of two-dimensional quantum heisenberg antiferromagnets with a nearly critical ground state, Phys. Rev. B 49, 11919 (1994).
  • Larsen and Lenoir [1988] J. Larsen and C. Lenoir, Synthesis of bis(ethylenedithio)tetrathiafulvalene (BEDT-TTF), Synthesis 1988, 134 (1988).
  • Abragam [1983] A. Abragam, The Principles of Nuclear Magnetism (Clarendon Press, Oxford, 1983).
  • Shimizu et al. [2006] Y. Shimizu, K. Miyagawa, K. Kanoda, M. Maesato, and G. Saito, Emergence of inhomogeneous moments from spin liquid in the triangular-lattice mott insulator κ\kappa-(ET)2Cu2(CN)3Phys. Rev. B 73, 140407 (2006).
  • Smith et al. [2003] D. F. Smith, S. M. De Soto, C. P. Slichter, J. A. Schlueter, A. M. Kini, and R. G. Daugherty, Dzialoshinskii-moriya interaction in the organic superconductor κ\kappa-(BEDT-TTF)2Cu[N(CN)2]Cl, http://link.aps.org/doi/10.1103/PhysRevB.68.024512,, Phys. Rev. B 68, 024512 (2003).
  • Mizusaki and Imada [2006] T. Mizusaki and M. Imada, Gapless quantum spin liquid, stripe, and antiferromagnetic phases in frustrated hubbard models in two dimensions, Phys. Rev. B 74, 014421 (2006).
  • Ghorbani et al. [2016] E. Ghorbani, L. F. Tocchio, and F. Becca, Variational wave functions for the S=12{S}=\frac{1}{2} Heisenberg model on the anisotropic triangular lattice: Spin liquids and spiral orders, Phys. Rev. B 93, 085111 (2016).
  • Takigawa et al. [1996] M. Takigawa, N. Motoyama, H. Eisaki, and S. Uchida, Dynamics in the S=1/2\mathit{S}\phantom{\rule{0.0pt}{0.0pt}}=\phantom{\rule{0.0pt}{0.0pt}}1/2 one-dimensional antiferromagnet sr2cuo3{\mathrm{sr}}_{2}{\mathrm{cuo}}_{3} via Cu63{}^{63}\mathrm{Cu} NMR, Phys. Rev. Lett. 76, 4612 (1996).
  • Brown et al. [1998] S. E. Brown, W. G. Clark, F. Zamborszky, B. J. Klemme, G. Kriza, B. Alavi, C. Merlic, P. Kuhns, and W. Moulton, 13C NMR measurements of the high-magnetic-field, low-temperature phases of (TMTTF)2PF6Phys. Rev. Lett. 80, 5429 (1998).
  • Cross [1979] M. C. Cross, Effect of magnetic fields on a spin-peierls transition, Phys. Rev. B 20, 4606 (1979).
  • Inagaki and Fukuyama [1984] S. Inagaki and H. Fukuyama, Spin-Peierls state vs antiferromagnetic state. III. effects of magnetic field, J. Phys. Soc. Jpn. 53, 4386 (1984).
  • Hase et al. [1993] M. Hase, I. Terasaki, K. Uchinokura, M. Tokunaga, N. Miura, and H. Obara, Magnetic phase diagram of the spin-Peierls cuprate CuGeO3Phys. Rev. B 48, 9616 (1993).
  • Brown et al. [1999] S. E. Brown, W. G. Clark, B. Alavi, D. Hall, M. J. Naughton, D. J. Tantillo, and C. A. Merlic, High-field magnetization of the spin-Peierls compound (TMTTF)2PF6Phys. Rev. B 60, 6270 (1999).
  • Hayashi and Ogata [2007b] Y. Hayashi and M. Ogata, Possibility of gapless spin liquid state by one-dimensionalization, J. Phys. Soc. Jpn. 76, 053705 (2007b).
  • Kaneko et al. [2014] R. Kaneko, S. Morita, and M. Imada, Gapless spin-liquid phase in an extended spin 1/2 triangular heisenberg model, J. Phys. Soc. Jpn. 83, 093707 (2014).
  • Ito et al. [2016] H. Ito, T. Asai, Y. Shimizu, H. Hayama, Y. Yoshida, and G. Saito, Pressure-induced superconductivity in the antiferromagnet κ\kappa-(ET)2CF3SO3 with quasi-one-dimensional triangular spin lattice, Phys. Rev. B 94, 020503 (2016).