WorksPhysics and Astronomy
Anomalies in the thermal conductivity of honeycomb antiferromagnet MnPS
Abstract
Intrinsic two-dimensional magnets serve as a good platform to explore collective, charge-neutral and low-energy excitations. Distinguishing the crucial role of them in experimental aspect remains a challenge for decades. Here, we study the thermal transport in honeycomb antiferromagnet MnPS with =78 K down to very low temperatures (<0.01). At high temperatures (>0.1), the field dependence of the thermal Hall conductivity exhibits a linear phonon Hall effect and a peak associated with the spin-flop transition due to a strong spin-lattice coupling, well reproducing the previous report (Phys. Rev. B 110, 165147 (2024)). Notably, below 2 K, we find that the field dependence of the thermal Hall conductivity exhibits sign reversals within the spin-flop phase, at which the field dependence of the longitudinal thermal conductivity also shows multiple valleys. We suggest that these anomalies are caused by the redistribution of Berry curvature in magnon bands, demonstrating the superior performance of the thermal Hall measurements to detect the Berry curvature distributions in magnetic insulators.
*Contact authors: [email protected] and [email protected]
Introduction
In recent years, the exploration of collective, charge-neutral, low-energy excitations in magnetic insulators has attracted considerable interest due to their potential for innovative approaches to manipulate and control thermal energy and information [1, 2]. To study these charge-neutral excitations in magnetic insulators, a thermal Hall effect (THE), which is a thermal analog of electrical Hall effects, has been pointed out as a powerful tool [3, 4, 5, 6]. Indeed, thermal Hall measurements have been adopted to study various charge-neutral excitations including magnons in ordered magnets [7, 8, 9, 10, 11], phonons [12, 13, 14, 15, 16, 17, 18], spinons in kagome antiferromagnets [19, 20, 21, 22], and Majorana fermions in Kitaev candidate materials [23, 24, 25, 26, 27, 28]. Although topological THEs by magnetic skyrmions have been studied in ferromagnetic and antiferromagnetic (AFM) three-dimensional magnets [9, 10], the correspondence between the experiments and the theoretical calculations has yet to be clarified because of the complexity of the spin Hamiltonian of these three-dimensional materials.
The two-dimensional (2D) van der Waals magnets provide an ideal platform not only to understand the mechanism of low-dimensional magnetism [29, 30, 31, 32] but also to study the topological magnon transports [33, 34] because of its good two dimensionality. Among these 2D magnets, the family of transition metal thiophosphates ( and ) with atoms forming a honeycomb lattice stacked by van der Waals interaction (Figure 1(a)) has been pointed out to be suited to study the two-dimensional magnetism because these compounds are known to exhibit Heisenberg-type, XY or XXZ-type, and Ising-type antiferromagnetism for and , respectively [35, 36, 37, 38]. In , which has the monoclinic crystal structure with the space group C2/m, the magnetic moments of ions () align antiferromagnetically along the direction tilted from the axis to the axis below the Néel temperature of K due to the AFM spin interactions [39, 40, 41, 42, 43, 44]. Due to this tilting, a spin-flop (SF) transition is induced by applying the magnetic field along the axis, as described in Figure 1(b). This noncollinear magnetic order induced by the SF transition is pointed out to realize not only a change in the longitudinal thermal transport through its large spin-lattice coupling [45, 46, 47], but also a finite THE that change drastically at the transition point [34, 48] even without the Dzyaloshinskii-Moriya interaction in this material [44]. Previous THE measurements of done above 20 K and below 9 T have succeeded in detecting finite thermal Hall conductivity , from which both an exchange striction and a single-ion magnetostriction are discussed to play important roles in the THE of [49].

Figure 1. (a) The side view of MnPS crystal structure with van der Waals gap (top) and the top view of the Mn honeycomb layer (bottom). (b) Schematic of the spin configurations across the spin-flop transition field . (c) The thermal-transport setup for measuring the longitudinal () and the transverse () thermal gradients simultaneously as a function of the applied heat current . (d) Temperature dependence of the magnetic susceptibility () measured at 1 T () after zero-field cooling (ZFC, the dashed line) and field cooling (FC, the solid line). The inside shows the field dependence of the magnetization () at different temperatures. The spin-flop field T is determined by the peak in the field derivative of as shown in Appendix § 9. (e) The temperature dependence of the longitudinal thermal conductivity () at 0 T (circles) and 15 T (squares). The temperature derivative of is shown on the right axis.
In this study, we extend the thermal-transport measurements of to lower temperatures () and higher magnetic fields. We find that the field dependence of the longitudinal thermal conductivity () below 50 K shows a large field suppression inside the SF phase, in addition to a sharp dip at the SF transition, showing a large contribution of the magnon in . Notably, below 2 K, we find that the field dependence of the thermal Hall conductivity () shows multiple positive and negative peaks in the SF phase, which exhibits a sign reversal as lowering the temperature. Simultaneously, the field dependence of shows multiple valleys at the fields corresponding to the peaks in . We find that these results are consistent with the theoretical calculations showing topological transitions of the magnon bands by the magnon band inversions [34] or multiple band crossings by the magnon-phonon hybridizations [48, 49]. We suggest that these anomalies observed in the field dependence of and in the SF phase at lower temperatures are caused by a redistribution of Berry curvatures in the magnon bands.
Materials and Methods
single crystals used in this work were grown by the chemical vapor transport method using as the transport agent. Manganese powder (99.98%, Alfa Aesar), red phosphorus powder (99.999%, Alfa Aesar), and sulfur powder (99.999%, Alfa Aesar) with a mole ratio of 1:1:3.05, and 0.15 g of as transport agent were weighed and loaded into a silicon quartz tube in an argon atmosphere, which was sealed under a high vacuum. 5 wt% extra sulfur was added for diminishing the vacancy in sulfur sites. The sealed quartz tube was placed in a two-zone tube furnace with the hot side of approximately 680 °C and the cold side of 630 °C. The whole growth process lasts approximately 7 days and then slowly cooled down to room temperature at a rate of 100 °C/h. Finally, lots of hexagonal green single crystal with the largest size of were obtained [50]. See Appendix § 7 for the X-ray diffraction data and pictures of our samples.
The magnetization () of the samples was measured using MPMS-XL5 (Quantum Design) above 2 K and a home-built Faraday-force magnetometer below 2 K by applying the magnetic field along the axis. The thermal-transport measurements were performed by the standard steady-state method by using one-heater-three-thermometers setup in a variable temperature insert for 2 K to 100 K and a dilution refrigerator below 2 K. The heat current and the magnetic field was applied parallel to the () and () axis of the sample, respectively [Fig. 1(c)]. Both the longitudinal () and the transverse () thermal differences were simultaneously measured to determine and (see Appendix § 8 for details).
Results and Discussions
As shown in Figure 1(d), the temperature dependence of the magnetic susceptibility () exhibits a broad maximum at around 100 K, which is followed by a sharp cusp at the Néel temperature K of without discernible difference between the field cooling and the zero-field cooling, reproducing previous results [32, 35, 46, 49]. The magnetic field dependence of the magnetization () at low temperatures shows a clear kink at the SF transition field ( T) that is determined by the peak of the field derivative of () as shown in Figure 8 in Appendix § 9. This kink in the field dependence of at blurs with increasing temperatures. Above 50 K, only a linear field dependence of is observed.

Figure 8. the field derivative of () at selected temperatures.
Figure 1(e) shows the temperature dependence of under 0 T and 15 T (). As shown in Figure 1(e), the temperature dependence of at both 0 T and 15 T exhibits a single peak at around 20 K, which is a general temperature dependence of a phonon thermal conductivity resulting from the competition between the increase of the phonon mean free path and the decrease of the phonon heat capacity as lowering the temperature [51]. The temperature derivative of displays a clear kink at , indicating a change in the phonon scatterings by the magnetic order through the strong spin-lattice coupling in , consistent with the drastic shift of Raman peak below [46]. A clear suppression of by applying 15 T along the axis is observed below around 50 K due to a suppression of a magnetic component in as discussed below. Below 1 K, shows a temperature dependence of with the mean free path comparable with the thickness of the sample, showing a ballistic phonon conduction observed in a high-quality sample (see Figure 9 in Appendix § 10).
Figure 9. (a) A log-log plot of the temperature dependence of of sample #2 measured in the dilution refrigerator. (b) The temperature dependence of the phonon mean free path at 0 T.
Figure 2(a) shows the field dependence of the normalized magnetothermal conductivity defined as at selected temperatures. As shown in Figure 2(a), at 50 K shows a monotonical decrease up to 15 T. At lower temperature, starts to show a sharp decrease above in addition to the monotonical decrease, concomitant with the clear indication of kink in the field dependence of . We note that our data well reproduces the previous study reporting above 9 K and below 9 T [49]. At 2 K, shows a positive magnetothermal conductivity below , followed by a dip at showing a clear correlation between and the change of the magnetic state. Furthermore, at 2 K shows a field suppression above 6 T, which is followed by an increase above 11 T.

Figure 2. (a,b) The field dependence of (a) and (b) at different temperatures. The vertical dashed-line shows . The data are vertically shifted for clarity. The dash-dotted lines in (b) show the linear fit of the data to determine the slope of . (c) The temperature dependence of at 15 T (left) and the that of the slope determined by the linear fit of the field dependence of (right).
This field dependence of shows that of contains both the phonon () and the magnetic () components. It is known that usually increases under a magnetic field because the magnetic field suppresses the spin fluctuations that scatter the phonons, which is observed as the positive magnetothermal conductivity below at 2 K (Figure 2(a)) as well as the one observed above 100 K reported in the previous work [49]. A field suppression of is induced by additional scattering on phonons, which can be caused by a closing of a magnon gap or resonance scatterings [22, 52]. The former can explain the dip in at and at 2 K; The Zeeman splitting of the degenerated magnon bands of the up and down spins at zero field induces the zero-energy crossing of the lower energy band at . This closing of the magnon gap increases the low-energy magnons that scatter phonons through the magnon-phonon coupling [45, 46, 47]. On the other hand, the field suppression of above 6 T is inconsistent with either the closing of the magnon gap and the resonance scattering, because no further zero-energy touching of the magnon bands is expected in this material [34, 48] and the magnetic field giving the minimum of in the resonance scattering should be lower than 6 T and should decrease with decreasing temperature [22] (1.5 T and 3.7 T at 2 K and 5 K, respectively for [53]) whereas the minimum of is observed at around 11 T both at 2 K and 5 K. Therefore, the field suppression of above 6 T should be attributed to a decrease of by the field increase of the magnon gap in the SF phase, indicating in turn a considerable contribution of in at low fields. The field increase of above 12 T observed below 5 K shows that the increase of by the field suppression effect on the magnon-phonon scatterings exceeds the decrease of by the increase of the magnon gap at higher fields.
Having established both the phonon and the magnon contributions in , we now discuss the field dependence of . As shown in Figure 2(b), the field dependence of above 30 K shows a negative linear increase as . Below 20 K, the field dependence of starts to show peaks and valleys for in addition to the linear field dependence. This linear field dependence in can be attributed to a phonon thermal Hall effect. Indeed, as shown in Figure 2(c), the temperature dependence of , which is estimated by a linear fit of the data above 12 T (the dash-dotted lines in Figure 2(b)), well follows that of at 15 T, where becomes dominant. This linear scaling between the temperature dependence of and that of is known as a tell-tale signature of the phonon thermal Hall effect as reported in non-magnetic insulators [16, 17]. Therefore, to investigate the details of the magnetic contribution in of MnPS, we subtract this linear component from data to estimate .
The field dependence of the subtracted thermal Hall conductivity () is plotted in Figure 3. As shown in Figure 3(e) and Figure 3(f), no discernible deviation from the linear field dependence is observed at 30 K and 50 K, showing the dominant phonon contribution in . At 20 K, shows a positive peak at 6 T and a negative peak at 7.6 T, of which the magnetic field are denoted as and , respectively. These peaks become larger with decreasing the temperature down to 5 K. Remarkably, at 2 K, the sign of these peaks of at and are reversed and a new negative peak appears at around 11 T (denoted as ). A weak positive peak might also appear at . Given that a sign of magnon directly reflects the sign of the dominant Berry curvature of the magnon bands [3, 4, 5], the multiple sign reversals in the field dependence of indicate the reversals of the dominant Berry curvature of the magnon bands, suggesting multiple topological phase transitions within the SF phase [34]. We note that a similar sign reversal of observed in another antiferromagnet NaCoTeO is also pointed out resulting from a topological transition of magnons by a SF transition [54, 55]. We stress that, although the positive peak at is reported in the previous work [49], the additional peaks at and as well as its sign reversal at 2 K are new features revealed by our measurements done at lower temperatures.

Figure 3. (a–f) The field dependence of the subtracted at 2 K to 50 K. The vertical dashed line indicates .
To further confirm the origin of THE in MnPS, we focus on the ultralow temperature region down to , where the phonon contributions is suppressed followed by the law. Figure 4(a) plots the field dependence of measured below 2 K in a different sample (#2). As shown in Figure 4(a), the field dependence of at 2 K reproduces the data of sample #1, i.e. the dip at , which is estimated to be almost the same with that in sample #1 (Figure 10 in Appendix § 11), and the field suppression above 6 T followed by the increase above 8 T. Below 2 K, the dip at is split into two and a new minimum appears at . The field of the first dip at 3.7 T may correspond to the onset of the SF transition and the second dip at 4.8 T to the peak of . The anomaly at is also observed as the negative peak in of sample #2 at 2 K (Figure 4(b)), reproducing the negative peak of in sample #1. These multiple valleys in the field dependence of is similar to that observed in the Kitaev candidate -RuCl [56], which is pointed out to be caused by multiple field-induced phase transitions [57]. Although the field dependence of near the SF transition is in good agreement with the theoretical calculation of the phonon with a magnon-phonon hybridization [49], other features are not, indicating the presence of a magnetic contribution.

Figure 4. The field dependence of (a) and (b) measured in sample #2 at 0.5 K, 1.0 K, and 2.0 K. The data of sample #1 at 2 K in Figure 2 is also shown for reference. The data is vertically shifted for clarity. The vertical dashed line indicates .
Figure 10. (a–c) the temperature dependence of the magnetic susceptibility (a), the field dependence of the magnetization (b), and the temperature dependence of the longitudinal thermal conductivity
The field dependence of the magnetization of sample #2 was also measured at 0.5 K and 1 K up to 10 T (Figure 11 in Appendix § 12). As shown in Figure 11, except the kink at the SF transition, no obvious anomalies in are observed within our resolution. This absence of anomaly in at the fields at which valleys of or peaks in are observed shows the high sensitivity of and to slight changes in the magnon bands associated with magnetic transitions. The former detects changes of the low-lying magnon populations and its scatterings, and the latter changes in the distribution of Berry curvature. On the other hand, magnetization only reflects the response of spins, averaged in the whole Brillouin zone, to changes in the magnetic field. This different sensitivity to topological transitions is indeed predicted in the theoretical calculations [34].
Figure 11. The magnetic field dependence of the magnetization () of sample #2 at 0.5 K (black line) and 1 K (red line). The field derivative of the data () is also shown on the right axis.
Here, we discuss the origin of the multiple sign reversals of observed in the SF phase (Figure 3(a)) based on the relation between the magnon and the Berry curvature. As described in Ref. [3, 4], is given by a summation of the Berry curvature weighted by a function of Bose factor as
where is the Berry curvature, , is a polylogarithm function, and is the Bose distribution function. This equation indicates that field dependence of a magnetic reflects the redistribution of the Berry curvature as a function of the field. Since this weighting factor is a function that monotonically decreases with increasing energy, the Berry curvature of the lower energy band generally dominates for magnon bands with the opposite Berry curvatures in the upper and the lower energy bands [58]. Good agreements between the theoretical estimations of based on Eq. (1) and the experiments have been observed in various materials including the ferromagnetic skyrmion material [9] and the antiferromagnetic kagome [20].
There are several possible origins that give rise to a sign change of . One possible origin is a topological phase transition in the AFM phase. Indeed, it is pointed out in Ref. [34] that the AFM spins on the honeycomb lattice under the magnetic field exhibit two topological phase transitions within the SF phase by the inversion of the magnon bands near the Brillouin-zone edge, in addition to the magnetic phase transitions into the SF phase. These topological phase transitions are manifested by the sign reversal of , which are consistent with the multiple sign reversals observed in the field dependence of (Figure 3). Such a topological phase transition can also be caused by a hybridization between the magnon and the phonon bands, in which a hybridization between the magnons and the phonons by a magnon-phonon coupling is pointed out to make band crossing points with Berry curvatures [48, 49]. These band crossings, which may occur in the SF phase under the magnetic field, could also give rise to a sign reversal of due to a field-induced redistribution of the Berry curvature.
On the other hand, the sign reversal of at and from 5 K to 2 K is rather unexpected in Ref. [34]. Also, the critical fields and the temperature showing the sign reversal of predicted by Ref. [34] are much higher than the observed values of our experiment. These differences could result from a difference of the spin interaction parameters used in the calculations [34] from the actual ones in our samples. Indeed, the neutron scattering experiments [41, 43] have shown the presence of further-neighbor interactions that are not included in the calculations [34], which would easily change the energy scale of the magnon band structure. Moreover, the SF field obtained in the calculation () is smaller than that observed in our samples (4.5 T), implying a difference in the spin Hamiltonian. Therefore, it remains an important future work to confirm the observed sign reversals of in the calculations including the further-neighbor interactions.
The second possibility is that consists of two different origins (such as magnons and phonons) with opposite sign of . In this case, the sign of will change depending on which component becomes dominant. Indeed, the sign reversals of at 5 K (Figure 2(b)) are observed only near at where the positive magnetic component exceeds the negative linear background from the phonon contribution. As shown in Figure 3, this linear phonon background needs to be subtracted to show the sign reversals of the magnetic component . However, our measurements at 2 K (Figure 2(b)) clearly demonstrate the sign reversals of without subtracting the linear background, excluding this two-component scenario as the origin of the sign reversals of in the SF phase of .
Another possibility is a contribution of the Berry curvature of the higher-energy bands at high temperatures. At a high temperature comparable to the spin interaction energy, a sign of is not simply determined by the Chern number of the lowest magnon band as shown by the calculations done in a ferromagnetic kagome [58, 59]. In these cases, the contribution of the Berry curvature from higher-energy bands becomes larger at higher temperatures whereas those in lower-energy bands smear out by a thermal broadening effect, giving rise to a sign change of as a function of temperature or field at high temperatures. In fact, in the calculations considering a hybridization between the magnon and phonon bands [48, 49], sign reversals of of are shown to appear at around at higher temperatures due to the anisotropy of multiple band crossing points within the narrow energy and momentum range. On the other hand, these magnon-phonon couplings induce the sign reversals near , not in the SF phase as observed in our measurements (especially at around 9 T in Figure 3(a)). Moreover, the magnitude of calculated by this hybridization is orders of magnitude smaller than that observed by our measurements. Therefore, it also remains a challenge for future research to try to reproduce observed in by introducing a magnon-phonon hybridization.
In summary, we have studied the thermal-transport properties in down to very low temperatures () and find that the field dependence of the thermal Hall conductivity exhibits multiple sign reversals above the SF transition field. Further, we find the field dependence of the longitudinal thermal conductivity also shows multiple valleys at the lowest temperature. We suggest that these anomalies are caused by the redistribution of Berry curvature in magnon bands, demonstrating the superior performance of the thermal Hall measurements to study the Berry curvature distributions in magnetic insulators.
Acknowledgments
We acknowledge Hyun-Yong Lee, Je-Geun Park, and Heejun Yang for insightful discussions. J.Y. was supported by Grant-in-Aid for JSPS Fellows. J.Y. and M.Y. thank the support by JSPS KAKENHI Grant Numbers JP22KF0111 and JP23H01116. X.L. and Y.P.S. was supported by the National Key R&D Program of China (Grants No. 2023YFA1607402 and No. 2021YFA1600201), the National Natural Science Foundation of China (Grants No. 12274412, No. 12204487 and No. U2032215), the CASHIPS Director’s Fund (Grant No. YZJJ2023QN34), and the Systematic Fundamental Research Program Leveraging Major Scientific and Technological Infrastructure, Chinese Academy of Sciences, under Contract No. JZHKYPT-437 2021-08.
Data availability
All the data supporting this study is available from the corresponding authors upon reasonable request.
Figures
X-ray diffraction and single crystal pictures
The X-ray diffraction data clearly show peaks, indicating that the surface of the crystals is -plane [Figure 5(a)]. We exfoliated MnPS with Scotch tape and chose two hexagonal shape samples with different thickness for the thermal-transport measurements. The sizes of two samples are about [#1, Figure 5(b)] and [#2, Figure 5(c)]. We note that there is a large ambiguity in the estimation of the thickness of sample #1 because of the thinness, which results in an ambiguity in the estimation of the magnitude of of sample #1 up to around 30%.

Figure 5. (a) X-ray diffraction data of the MnPS single crystal. (b,c) Pictures of the thermal-transport setup for sample #1 (b) and sample #2 (c).
The setup of the thermal-transport measurements
For our thermal-transport measurements, One heater and three thermometers (Cernox™ (CX-1050) thermometers for measurements above 2 K and RuO thermometers below 2 K) were attached to the sample via gold wires so that both the longitudinal () and the transverse () thermal differences can be measured simultaneously. These resistive thermometers have the advantage of the high sensitivity at low temperatures. All thermometers were carefully calibrated in magnetic fields by a calibrated thermometer placed outside the magnetic field. The heat current was applied parallel to the () plane of the sample and the magnetic field was applied along the () axis, where is the heater power, the width of the sample between the thermal contacts to read and , and the thickness of the sample. To cancel the longitudinal component in due to a misalignment of the thermal contacts, was asymmetrized with respect to the field direction as .
The longitudinal thermal conductivity and thermal Hall conductivity is derived by
where is the length between and [60]. In this estimation, the anisotropy in the – plane due to the monoclinic crystal structure of this sample is ignored (i. e. ) because the distortion from a perfect hexagonal lattice is very small; the axis length is close to expected for the hexagonal lattice [39]. The small magnetic anisotropy in the – plane has also been shown in the neutron scattering experiments [43].
The direction of the heat current was confirmed to be along the axis for sample #2 via the XRD measurements done after the thermal-transport measurements. Unfortunately, the direction of the axis of sample #1 could not be determined because the sample shattered into pieces when it was removed from the thermal-transport measurement cell after the measurements. Assuming the relationship between the hexagonal sample shape and the crystal axis remains the same with sample #2, it is likely that the heat current was also applied along the axis in sample #1.
According to Eq. (2), is expected to linearly depend on . We checked this linear relation by measuring the field dependence of at different . As shown in Figure 6, increases linearly with (the inset of Figure 6(a)) and the field dependence of determined by measured with the different shows a good match, confirming the validity of our thermal Hall measurements.

Figure 6. (a) The field dependence of measured by applying different . The inset shows measured at 15 T as a function of . (b) The field dependence of determined from shown in (a).
Figure 7 shows the raw data of the temperatures during the measurements at 20 K shown in Figure 6. The field dependence of is obtained by antisymmetrizing the data averaged for about 6 minutes before changing the heater status. The standard error of the data is typically about 0.020 mK, which is about the same size as the symbols in Figure 6 (a).

Figure 7. (a) A raw time chart of the sample temperatures ( black line, yellow line, and cyan line) during the field sweep (red line, shown in the right axis) measurements at 20 K. (b) An enlarged figure of the first five hours data. The heater of 1.98 mW was turned on and off
at the time indicated by the red and black dashed line, respectively. The average temperature of and was kept constant during the measurements. (c) A time chart of the transverse temperature difference (blue line) obtained from the data shown in (b).
Determination of
The spin-flop transition field () is determined by the peak of the field derivative of the magnetization () as shown in Figure 8.
Estimation of the phonon mean free path
Fig. 9(a) shows the temperature dependence of below 3 K. The power law fitting of the data () below 1 K gives the exponent of temperature dependence , which is close to that of boundary-limited phonons (). In fact, the long mean free path of phonons comparable to the sample size is obtained as following. The thermal conductivity of phonons is given by
where , , and is the heat capacity, the sound velocity, and the mean free path of phonons, respectively. We estimate by a fitting to the heat capacity data [61] as J mol K and the molar volume of m mol. The sound velocity is estimated from the phonon dispersion of [49] as 3454 m/s. Assuming the dominant phonon contribution in below 1 K, the temperature dependence of the phonon mean free path (Figure 9(b)) is obtained from the temperature dependence of (Figure 9(a)). As shown in Figure 9(b), exceeds the thickness of the sample below 3 K, demonstrating the ballistic phonon conduction observed in high-quality samples. The increase of beyond the thickness of the sample at lower temperature could be due to specular scatterings that is known to enhances beyond the sample thickness, or an additional magnon contribution in .
The sample dependence of the magnetization and the thermal conductivity
We checked the reproducibility of our data by measuring two samples (#1 and #2). As shown in Figure 10, the temperature dependence of the magnetic susceptibility (Figure 10(a)), the field dependence of the magnetization (Figure 10(b)), and the temperature dependence of the longitudinal thermal conductivity (Figure 10(c)) of the two samples are very similar, showing the good reproducibility.
On the other hand, differences in the field dependence of and at 2 K are observed as shown in Figure 4; the field dependence of of sample #2 at 2 K shows a larger minimum at a lower field of 8 T than that in sample #1, and the positive peak of of sample #1 at is not observed in sample #2.
One possible reason for this sample dependence is the difference in the sample thickness; the sample #2 ( mm) is about three times thicker than that of sample #1 ( mm). Therefore, of sample #2 can be longer than that of sample #1, as shown by the larger of sample #2 (Figure 10(c)). This difference in the phonon mean free path would result in a different relative magnitude of to , and the different field dependence of due to a different field dependence of and .
As shown in Figure 10 (a), the magnetic susceptibility of sample #2 shows a larger Curie contribution than that in sample #1 below around 15 K, implying that sample #2 contains more magnetic impurities than that in sample #1. This difference in the amount of the magnetic impurities might also affect the field dependence of and . It should be noted that the magnitude of of this compound is not a good measure of the sample quality. As shown in Figure 10 (c), the clear phonon peak is observed in the temperature dependence of at around 20 K in both samples, showing that becomes comparable to the sample size, much longer than the impurity size. Thus, is no longer suppressed by these impurities.
In addition, this sample dependence might be brought by a different degree of distortions within the sample (e.g. stacking faults). Given the high sensitivity of the van der Waals structure to a mechanical stress, some distortions might be easily introduced during the sample growth or the setup for the thermal-transport measurements. Indeed, a previous transmission electron microscopy [62] confirms a large population of the stacking faults in (), which is pointed out to reduce the cross-plane thermal conductivity of the atomically thin samples. These impurities or distortions might affect the spin interaction parameters as well as the coupling constant between the spins and the lattice, which could alter the field dependence of as pointed out in Ref. [34, 49]. A similar effect related to the sample quality is also pointed out in the van der Waals magnet -RuCl (see Ref. [27, 63] for example).
at T (c) of two MnPS samples (#1 and #2). The inset in (b) shows the field dependence of the field derivative of the magnetization.
During the preparation of this manuscript, we became aware of a similar study on MnPS down to K () that reports a broad sample dependence with a very different magnetic field dependence of with a much larger magnitude from both the previous work [49] and our results, which points out the importance of a magnetic contribution beyond a phonon contribution from a magnon-phonon hybridization [61]. Although the sign reversals in the field dependence of might be reproduced by separating the phonon contribution in the data reported in Ref. [61], the difference of the two-orders of magnitude in from our data and the previous work [49] would be difficult to explain even by considering a magnon-phonon hybridization. This is because, in the formalism of the intrinsic thermal Hall effect (Eq. (1)), is given by a summation of the Berry curvature weighted by a function of Bose factor [3, 4]. Given that a finite Berry curvature is caused by a secondary interaction (such as Dzyaloshinskii–Moriya and spin–lattice coupling) and the Nielsen–Ninomiya theorem regarding the Berry curvature in a magnetic insulator, the magnitude of the Berry curvature of each band cannot be extremely large, limiting the order of magnitude of from this intrinsic Berry-curvature expression to where is the interlayer distance. The magnitude of this is roughly on the order of mW K m at K, which is in the same order with our results and that from Ref. [49]. In fact, this size of observed in the ferromagnetic skyrmion material [9] and in the antiferromagnetic kagome [20] have been shown to agree with the theoretical calculations by the intrinsic Berry-curvature expression (Eq. (1)) with respect to not only the qualitative temperature dependence of but also their quantitative magnitude.
Therefore, to explain the enhancement of by the two-orders magnitude, one needs to consider a different origin distinct from the intrinsic Berry-curvature expression. One possibility is an extrinsic origin. As known in the anomalous Hall effect in ferromagnetic metals, the extrinsic skew contribution exceeds the intrinsic contribution in a sample with a larger electric conductivity (see Ref. [64] for example). Indeed, it has been demonstrated that a similar extrinsic mechanism appears as a spin thermal Hall effect in magnetic insulators [21] for samples with larger . However, given that the magnitude of only differs a factor of or in the data reported in Ref. [61] from our results, it might also be questionable if this can explain the two-orders difference in the magnitude of .
The magnetization data below 2 K in sample #2
Figure 11 shows the field dependence of of sample #2 measured at 0.5 K and 1.0 K by using our home-built Faraday-force magnetometer installed in a dilution refrigerator. As shown in Figure 11, no obvious anomaly is observed in the magnetization curve within our resolution, except the kink at the spin-flop transition at 4.5 T.
References
- [1]M. Maldovan, Sound and heat revolutions in phononics, Nature 503, 209 (2013).DOI
- [2]A. V. Chumak, V. I. Vasyuchka, A. A. Serga, and B. Hillebrands, Magnon spintronics, Nat. Phys. 11, 453 (2015).
- [3]H. Katsura, N. Nagaosa, and P. A. Lee, Theory of the Thermal Hall Effect in Quantum Magnets, Phys. Rev. Lett. 104, 066403 (2010).
- [4]R. Matsumoto, R. Shindou, and S. Murakami, Thermal Hall effect of magnons in magnets with dipolar interaction, Phys. Rev. B 89, 054420 (2014).
- [5]Y.-F. Yang, G.-M. Zhang, and F.-C. Zhang, Universal Behavior of the Thermal Hall Conductivity, Phys. Rev. Lett. 124, 186602 (2020).
- [6]L. Zhang, J. Ren, J.-S. Wang, and B. Li, Topological magnon insulator in insulating ferromagnet, Phys. Rev. B 87, 144101 (2013).
- [7]Y. Onose, T. Ideue, H. Katsura, Y. Shiomi, N. Nagaosa, and Y. Tokura, Observation of the magnon Hall effect, Science 329, 297 (2010).
- [8]T. Ideue, Y. Onose, H. Katsura, Y. Shiomi, S. Ishiwata, N. Nagaosa, and Y. Tokura, Effect of lattice geometry on magnon Hall effect in ferromagnetic insulators, Phys. Rev. B 85, 134411 (2012).DOI
- [9]M. Akazawa, H.-Y. Lee, H. Takeda, Y. Fujima, Y. Tokunaga, T.-h. Arima, J. H. Han, and M. Yamashita, Topological thermal Hall effect of magnons in magnetic skyrmion lattice, Phys. Rev. Res. 4, 043085 (2022).
- [10]H. Takeda, M. Kawano, K. Tamura, M. Akazawa, J. Yan, T. Waki, H. Nakamura, K. Sato, Y. Narumi, M. Hagiwara et al., Magnon thermal Hall effect via emergent SU(3) flux on the antiferromagnetic skyrmion lattice, Nat. Commun. 15, 566 (2024).DOI
- [11]H. Zhang, C. Xu, C. Carnahan, M. Sretovenic, N. Suri, D. Xiao, and X. Ke, Anomalous thermal Hall effect in an insulating van der waals magnet, Phys. Rev. Lett. 127, 247202 (2021).
- [12]K. Sugii, M. Shimozawa, D. Watanabe, Y. Suzuki, M. Halim, M. Kimata, Y. Matsumoto, S. Nakatsuji, and M. Yamashita, Thermal Hall Effect in a Phonon-Glass Ba₃CuSb₂O₉, Phys. Rev. Lett. 118, 145902 (2017).
- [13]G. Grissonnanche, S. Thériault, A. Gourgout, M. E. Boulanger, E. Lefrançois, A. Ataei, F. Laliberté, M. Dion, J. S. Zhou, S. Pyon et al., Chiral phonons in the pseudogap phase of cuprates, Nat. Phys. 16, 1108 (2020).
- [14]M.-E. Boulanger, G. Grissonnanche, S. Badoux, A. Allaire, É. Lefrançois, A. Legros, A. Gourgout, M. Dion, C. H. Wang, X. H. Chen et al., Thermal Hall conductivity in the cuprate Mott insulators Nd₂CuO₄ and Sr₂CuO₂Cl₂, Nat. Commun. 11, 5325 (2020).DOI
- [15]T. Uehara, T. Ohtsuki, M. Udagawa, S. Nakatsuji, and Y. Machida, Phonon thermal Hall effect in a metallic spin ice, Nat. Commun. 13, 4604 (2022).DOI
- [16]X. Li, B. Fauqué, Z. Zhu, and K. Behnia, Phonon thermal Hall effect in strontium titanate, Phys. Rev. Lett. 124, 105901 (2020).
- [17]X. Li, Y. Machida, A. Subedi, Z. Zhu, L. Li, and K. Behnia, The phonon thermal Hall angle in black phosphorus, Nat. Commun. 14, 1027 (2023).
- [18]L. Chen, L. Le Roux, G. Grissonnanche, M.-E. Boulanger, S. Thériault, R. Liang, D. A. Bonn, W. N. Hardy, S. Pyon, T. Takayama et al., Planar Thermal Hall Effect from Phonons in Cuprates, Phys. Rev. X 14, 041011 (2024).
- [19]M. Hirschberger, R. Chisnell, Y. S. Lee, and N. P. Ong, Thermal Hall effect of spin excitations in a kagome magnet, Phys. Rev. Lett. 115, 106603 (2015).
- [20]H. Doki, M. Akazawa, H.-Y. Lee, J. H. Han, K. Sugii, M. Shimozawa, N. Kawashima, M. Oda, H. Yoshida, and M. Yamashita, Spin Thermal Hall Conductivity of a Kagome Antiferromagnet, Phys. Rev. Lett. 121, 097203 (2018).DOI
- [21]M. Akazawa, M. Shimozawa, S. Kittaka, T. Sakakibara, R. Okuma, Z. Hiroi, H.-Y. Lee, N. Kawashima, J. H. Han, and M. Yamashita, Thermal Hall Effects of Spins and Phonons in Kagome Antiferromagnet Cd-Kapellasite, Phys. Rev. X 10, 041059 (2020).DOI
- [22]D. Watanabe, K. Sugii, M. Shimozawa, Y. Suzuki, T. Yajima, H. Ishikawa, Z. Hiroi, T. Shibauchi, Y. Matsuda, and M. Yamashita, Emergence of nontrivial magnetic excitations in a spin-liquid state of kagomé volborthite, Proc. Natl Acad. Sci. USA 113, 8653 (2016).DOI
- [23]Y. Kasahara, T. Ohnishi, Y. Mizukami, O. Tanaka, S. Ma, K. Sugii, N. Kurita, H. Tanaka, J. Nasu, Y. Motome et al., Majorana quantization and half-integer thermal quantum Hall effect in a Kitaev spin liquid, Nature 559, 227 (2018).DOI
- [24]M. Yamashita, J. Gouchi, Y. Uwatoko, N. Kurita, and H. Tanaka, Sample dependence of half-integer quantized thermal Hall effect in the Kitaev spin-liquid candidate α-RuCl3, Phys. Rev. B 102, 220404 (2020).DOI
- [25]T. Yokoi, S. Ma, Y. Kasahara, S. Kasahara, T. Shibauchi, N. Kurita, H. Tanaka, J. Nasu, Y. Motome, C. Hickey et al., Half-integer quantized anomalous thermal Hall effect in the Kitaev material candidate α-RuCl3, Science 373, 568 (2021).DOI
- [26]K. Imamura, S. Suetsugu, Y. Mizukami, Y. Yoshida, K. Hashimoto, K. Ohtsuka, Y. Kasahara, N. Kurita, H. Tanaka, P. Noh et al., Majorana-fermion origin of the planar thermal Hall effect in the Kitaev magnet α-RuCl3, Sci. Adv. 10, eadk3539 (2024).DOI
- [27]J. A. N. Bruin, R. R. Claus, W. Matsumoto, N. Kurita, H. Tanaka, and H. Takagi, Robustness of the thermal Hall effect close to half-quantization in α-RuCl3, Nat. Phys. 18, 401 (2022).DOI
- [28]Y. Matsuda, T. Shibauchi, and H.-Y. Kee, Kitaev quantum spin liquids, Rev. Mod. Phys. 97, 045003 (2025).
- [29]B. Huang, G. Clark, E. Navarro-Moratalla, D. R. Klein, R. Cheng, K. L. Seyler, D. Zhong, E. Schmidgall, M. A. McGuire, and D. H. Cobden, Layer-dependent ferromagnetism in a van der Waals crystal down to the monolayer limit, Nature 546, 270 (2017).DOI
- [30]C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, W. Bao, C. Wang, and Y. Wang, Discovery of intrinsic ferromagnetism in two-dimensional van der Waals crystals, Nature 546, 265 (2017).DOI
- [31]Z. Fei, B. Huang, P. Malinowski, W. Wang, T. Song, J. Sanchez, W. Yao, D. Xiao, X. Zhu, and A. F. May, Two-dimensional itinerant ferromagnetism in atomically thin Fe3GeTe2, Nat. Mater. 17, 778 (2018).DOI
- [32]G. Long, H. Henck, M. Gibertini, D. Dumcenco, Z. Wang, T. Taniguchi, K. Watanabe, E. Giannini, and A. F. Morpurgo, Persistence of magnetism in atomically thin MnPS3 crystals, Nano Lett. 20, 2452 (2020).DOI
- [33]S. A. Owerre, A first theoretical realization of honeycomb topological magnon insulator, J. Phys.: Condens. Matter 28, 38600 (2016).DOI
- [34]R. R. Neumann, A. Mook, J. Henk, and I. Mertig, Thermal Hall Effect of Magnons in Collinear Antiferromagnetic Insulators: Signatures of Magnetic and Topological Phase Transitions, Phys. Rev. Lett. 128, 117201 (2022).
- [35]P. Joy and S. Vasudevan, Magnetism in the layered transition-metal thiophosphates MPS3 (M = Mn, Fe, and Ni), Phys. Rev. B 46, 5425 (1992).
- [36]A. R. Wildes, V. Simonet, E. Ressouche, G. J. McIntyre, M. Avdeev, E. Suard, S. A. Kimber, D. Lancon, G. Pepe, and B. Moubarki, Magnetic structure of the quasi-two-dimensional antiferromagnet NiPS3, Phys. Rev. B 92, 224408 (2015).
- [37]K. C. Rule, G. McIntyre, S. J. Kennedy, and T. J. Hicks, Single-crystal and powder neutron diffraction experiments on FePS3: Search for the magnetic structure, Phys. Rev. B 76, 134402 (2007).DOI
- [38]D. Lancon, H. Walker, E. Ressouche, B. Ouladdiaf, K. Rule, G. McIntyre, T. Hicks, H. M. Rønnow, and A. Wildes, Magnetic structure and magnon dynamics of the quasi-two-dimensional antiferromagnet FePS3, Phys. Rev. B 94, 214407 (2016).
- [39]K. Kurosawa, S. Saito, and Y. Yamaguchi, Neutron diffraction study on MnPS₃ and FePS₃, J. Phys. Soc. Jpn. 52, 3919 (1983).
- [40]K. Okuda, K. Kurosawa, S. Saito, M. Honda, Z. Yu, and M. Date, Magnetic properties of layered compound MnPS₃, J. Phys. Soc. Jpn. 55, 4456 (1986).
- [41]A. Wildes, B. Roessli, B. Lebech, and K. Godfrey, Spin waves and the critical behaviour of the magnetization in MnPS₃, J. Phys.: Condens. Matter 10, 6417 (1998).
- [42]D. J. Goossens, A. Wildes, C. Ritter, and T. J. Hicks, Ordering and the nature of the spin flop phase transition in MnPS₃, J. Phys.: Condens. Matter 12, 1845 (2000).DOI
- [43]A. Wildes, H. M. Rønnow, B. Roessli, M. Harris, and K. Godfrey, Static and dynamic critical properties of the quasi-two-dimensional antiferromagnet MnPS₃, Phys. Rev. B 74, 094422 (2006).
- [44]A. R. Wildes, S. Okamoto, and D. Xiao, Search for nonreciprocal magnons in MnPS₃, Phys. Rev. B 103, 024424 (2021).DOI
- [45]Y.-J. Sun, Q.-H. Tan, X.-L. Liu, Y.-F. Gao, and J. Zhang, Probing the magnetic ordering of antiferromagnetic MnPS₃ by Raman spectroscopy, J. Phys. Chem. Lett. 10, 3087 (2019).
- [46]K. Kim, S. Y. Lim, J. Kim, J.-U. Lee, S. Lee, P. Kim, K. Park, S. Son, C.-H. Park, and J.-G. Park, Antiferromagnetic ordering in van der Waals 2D magnetic material MnPS₃ probed by Raman spectroscopy, 2D Mater. 6, 041001 (2019).
- [47]D. Vaclavkova, A. Delhomme, C. Faugeras, M. Potemski, A. Bogucki, J. Suffczyński, P. Kossacki, A. R. Wildes, B. Grémaud, and A. Saúl, Magnetoelastic interaction in the two-dimensional magnetic material MnPS₃ studied by first principles calculations and Raman experiments, 2D Mater. 7, 035030 (2020).
- [48]G. Go, H. Yang, J.-G. Park, and S. K. Kim, Topological magnon polarons in honeycomb antiferromagnets with spin-flop transition, Phys. Rev. B 109, 184435 (2024).
- [49]H. Yang, G. Go, J. Park, S. K. Kim, and J.-G. Park, Exchange striction induced thermal Hall effect in the van der Waals antiferromagnet MnPS₃, Phys. Rev. B 110, 165147 (2024).
- [50]L. Cheng, F. Formisano, K. A. Grishunin, S. D. Gorelov, P. H. M. van Loosdrecht, J. Yan, X. Luo, Z. Sheng, and E. A. Mashkovich, Nonlinear terahertz Kerr effect in quasi-2D MnPS₃, Opt. Lett. 47, 4052 (2022).
- [51]K. Behnia, Fundamentals of Thermoelectricity (Oxford University Press, 2015).
- [52]B. R., Resonant Scattering. Thermal Conduction in Solids (Oxford University Press, Oxford, 1976), p. 91-99.
- [53]J. J. Abraham, Y. Senyk, Y. Shemerliuk, S. Selter, S. Aswartham, B. Büchner, V. Kataev, and A. Alfonsov, Magnetic anisotropy and low-energy spin dynamics in the van der Waals compounds Mn₂P₂S₆ and MnNiP₂S₆, Phys. Rev. B 107, 165141 (2023).
- [54]H. Takeda, J. Mai, M. Akazawa, K. Tamura, J. Yan, K. Moovendran, K. Raju, R. Sankar, K.-Y. Choi, and M. Yamashita, Planar thermal Hall effects in the Kitaev spin liquid candidate Na₂Co₂TeO₆, Phys. Rev. Res. 4, L042035 (2022).
- [55]N. Li, R. R. Neumann, S. K. Guang, Q. Huang, J. Liu, K. Xia, X. Y. Yue, Y. Sun, Y. Y. Wang, Q. J. Li et al., Magnon-polaron driven thermal Hall effect in a Heisenberg-Kitaev antiferromagnet, Phys. Rev. B 108, L140402 (2023).
- [56]P. Czajka, T. Gao, M. Hirschberger, P. Lampen-Kelley, A. Banerjee, J. Yan, D. G. Mandrus, S. E. Nagler, and N. P. Ong, Oscillations of the thermal conductivity in the spin-liquid state of α-RuCl₃, Nat. Phys. 17, 915 (2021).
- [57]J. A. N. Bruin, R. R. Claus, Y. Matsumoto, J. Nuss, S. Laha, B. V. Lotsch, N. Kurita, H. Tanaka, and H. Takagi, Origin of oscillatory structures in the magnetothermal conductivity of the putative Kitaev magnet α-RuCl₃, APL Materials 10, 090703 (2022).
- [58]A. Mook, J. Henk, and I. Mertig, Magnon Hall effect and topology in kagome lattices: A theoretical investigation, Phys. Rev. B 89, 134409 (2014).
- [59]H. Lee, J. H. Han, and P. A. Lee, Thermal Hall effect of spins in a paramagnet, Phys. Rev. B 91, 125413 (2015).
- [60]M. Yamashita, M. Akazawa, M. Shimozawa, T. Shibauchi, Y. Matsuda, H. Ishikawa, T. Yajima, Z. Hiroi, M. Oda, H. Yoshida et al., Metal-transport studies of kagomé antiferromagnets, J. Phys.: Condens. Matter 32, 074001 (2020).
- [61]M. Nawwar, R. R. Neumann, J. Wen, I. Mertig, A. Mook, and J. P. Heremans, Large thermal Hall effect in MnPS₃, Rep. Prog. Phys. 88, 080503 (2025).
- [62]H. Ju, D.-G. Jeong, Y.-G. Choi, S. Son, W.-G. Jung, M.-C. Jung, S. Kang, M. J. Han, B.-J. Kim, J.-G. Park et al., Influence of stacking disorder on cross-plane thermal transport properties in TMPS₃ (TM = Mn, Ni, Fe), Appl. Phys. Lett. 117, 063103 (2020).
- [63]H. Zhang, M. A. McGuire, A. F. May, H.-Y. Chao, Q. Zheng, M. Chi, B. C. Sales, D. G. Mandrus, S. E. Nagler, H. Miao et al., Stacking disorder and thermal transport properties of α-RuCl₃, Phys. Rev. Mater. 8, 014402 (2024).
- [64]S. Onoda, N. Sugimoto, and N. Nagaosa, Intrinsic Versus Extrinsic Anomalous Hall Effect in Ferromagnets, Phys. Rev. Lett. 97, 126602 (2006).